Collins Configuration

Making Collins yours, a groovy kind of love


Things to know before you configure

Collins configuration files are type safe. That is, you can't create a configuration that will lead to a runtime error. If collins detects a configuration error it will fail to start. If you modify a config after the system starts and that config is invalid, it will not load and collins will complain loudly. If this is not the case, please file a bug.

Every type of configuration comes with a reference config. Reference configuration files can be found in the github repo. Reference files are not modified by users, they serve to provide Collins with default values as well as to tell it what the appropriate value types are (lists, strings, etc). Most reference configurations come with some level of documentation as well. In general you should use and make modifications to conf/application.conf.

The configuration file that you create for collins is overriding the values found in the reference configs. There is a sample configuration in the github repo for you to use as a starting point.

You can pass your configuration file to collins by using a system property. Pass -Dconfig.file=/path/to/file will cause that file to be used for the application.


Configuration options that don't fit anywhere else

Application Specific

These configuration aspects are specific to the play framework.

application.secret String A secret key used to secure your cookies. If you deploy several collins instances be sure to use the same key.
parsers.text.maxLength Integer The maximum allowable size of a body. Be sure to make this large enough to allow for LSHW/LLDP uploads.
evolutionplugin String Set to disabled in production, or enabled in development
evolutionplugin = disabled


The only crypto configuration available is the key. The crypto key must not be changed after the system is initialized for the first time or you will be unable to decrypt data.

crypto.key String A 63 character random string. Be sure that any Collins instances using the same database have the same crypto.key
crypto.key = "lkkajsdlkajsdlkajsdlkajsdlkajsdlkajsdlkajsdlkajsdlkajsdlkajsdla"


Collins is extensively tested using both MySQL 5.5 and H2. H2 is used for development and MySQL is used in production.
db {
  collins {
    logStatements = false
    user = "sa"
    password = ""
    driver = "org.h2.Driver"
    url = "jdbc:h2:mem:play;IGNOERCASE=true"
db {
  collins {
    logStatements = false
    user = "root"
    password = ""
    driver = "com.mysql.jdbc.Driver"
    url = "jdbc:mysql://localhost/collins?autoReconnect=true&interactiveClient=true"

LSHW Parsing

Due to how LSHW handles PCIe flash devices, you have to give the parser a hint as to which attribute values indicate that the device is in fact a flash storage device. Here you can specify the string to match on and the size to use if the match is found.

lshw.defaultNicCapacity Long When parsing NIC capicity collins tries a few things. It first uses the capacity value, if it is found. If not, it uses the size value, if it is found. If neither capacity or size are found, it then looks in the product name for the string '10-gig'. If it finds that string, it assumes the size is 10000000000 bits. If none of these succeed, it will use the lshw.defaultNicCapacity value. This configuration is specified in bits, and there is no default value. If this value is not set, and collins tries to use it, it will throw an exception and fail to parse the LSHW data.
lshw.flashProduct String Deprecated Product string to look for in LSHW output to detect a flash device. Defaults to flashmax. This is deprecated, and will be removed in a future version. Please use lshw.flashProducts instead.
lshw.flashProducts Set[String] Product strings to look for in LSHW output to detect a flash device. Defaults to ["flashmax"], and includes lshw.flashProduct.
lshw.flashSize Long If the flashProduct matched, the size to use. Defaults to 1400000000000.
lshw.includeDisabledCpu Bool Whether CPUs marked as disabled should be included. Defaults to false.
lshw.includeEmptySocket Bool Whether sockets with no CPU should be included. Defaults to false.
lshw {
  defaultNicCapacity = 10000000000
  flashProducts = ["flashmax","fusionIO","TachION Drive"]
  flashSize = 1400000000000

Query Log

The query log is primary used for development but allows you to turn on database query logging and API (find/search) logging.

querylog.enabled Boolean Whether to log database queries
querylog.includeResults Boolean Defaults to false. If the query log is enabled, and includeResults is true, query results will also be logged
querylog.prefix String Defaults to "QUERY: ". A prefix to use in the log message.
querylog.frontendLogging Boolean Defaults to false. If true, API queries are logged
querylog {
  enabled = false
  includeResults = false
  prefix = "QUERY: "
  frontendLogging = false


Setting up user authentication and authorization

Authentication Namespace authentication

Collins supports a variety of authentication backends along with a flexible API endpoint based permissions system. Some features also have specific permissions.

Collins comes bundled with a small in-memory authentication mechanism for testing, file based authentication for small or simple installations, and LDAP based authentication for large or enterprise installations. If using LDAP we recommend running an LDAP slave on the collins node itself in order to be resilient to large outages where the usual LDAP infrastructure might be unavailable.

type String Defaults to default, which is the in-memory database. Other valid values are file or ldap
permissionsFile String Path to the YAML file containing the application permissions. There is a sample permissions file below. Every API endpoint has a name associated with it. You can configure per endpoint permissions as well as some feature permissions.
permissionsCacheSpecification String The guava cache specification to use for the permissions file, as described in Defaults to "expireAfterWrite=30s".
adminGroup Set[String] A list of groups for administrative users. Users that have one of these groups will be in the admin group. Note that being in the adminGroup doesn't always mean you will have access to something, but by default it will. Using the permissionsFile it is possible to configure permissions such that even someone in the admin group can't access a partilar feature or API endpoint.

The permissions example below creates two user groups, admins and tools. User groups are just named groups of users that are stored in the YAML file. This is useful for people authenticating with flat files or if you want to store some group information here. The admins user group has two users, alan and bmatheny. The tools user group has three users. There are two feature permissions specified, one for rate limiting and the other for who can see encrypted tags, like passwords. The last permission is for the powerStatus API endpoint and is open to a broader group of users. Every API permission name is documented in the API documentation.

    - "alan"
    - "bmatheny"
    - "jetpants"
    - "itouch"
    - "phil"
    - "u=tools"
    - "u=admins"
    - "u=tools"
    - "g=infra"
    - "g=infra"
    - "g=ops"
    - "u=admins"
    - "u=tools"
authentication {
  type = default
  permissionsFile = "conf/permissions.yaml"
  permissionsCacheSpecification = "expireAfterWrite=30s"
  adminGroup = [Infra]

LDAP Authentication Namespace authentication.ldap

At Tumblr LDAP is the primary authentication mechanism. We have tested that this code works with at least RFC 2307, IPA, and the OS X admin server.

host String The LDAP host to authenticate against. This is a required configuration.
ssl Boolean Whether or not to use LDAPS when communicating with the LDAP server. Defaults to false
schema Enum One of rfc2307 or rfc2307bis. This is a required configuration.
searchbase String The LDAP search base to use. This is a required configuration.
usersub String Where to find users in the LDAP schema. This is a required configuration.
groupsub String Where to find groups in the LDAP schema. This is a required configuration.
anonymous Boolean Whether to use anonymous searches. Defaults to false
binddn String The bind DN to use for ldap searches. Only used when anonymous is set to false. Defaults to ""
bindpwd String The bind password to use for ldap searches. Only used when anonymous is set to false. Defaults to ""
userAttribute String The LDAP schema attribute used to identify the user name. Defaults to uid
userNumberAttribute String The LDAP schema attribute used to identify the user id number. Defaults to uidNumber
groupAttribute String An LDAP schema attribute to use for retrieving group information. This is a required configuration
groupNameAttribute String An LDAP schema attribute to use for identifying the group set in groupAttribute. Defaults to cn
cacheSpecification String The guava cache specification to use for LDAP, as described in Defaults to "expireAfterWrite=30s".
authentication {
  type = ldap
  ldap {
    host = "localhost"
    groupAttribute = "uniqueMember"
    searchbase = "dc=example,dc=org"
    usersub = "ou=people"
    groupsub = "ou=groups"
    cacheSpecification = "expireAfterWrite=30s"

File Authentication Namespace authentication.file

The file format used is a variation on that used by Apache. You can find a sample script for generating users and passwords in scripts/

userfile String The file to look in for auth information
cacheSpecification String The guava cache specification to use for file-based auth, as described in Defaults to "expireAfterWrite=30s".
authentication {
  type = file
  file {
    userfile = "conf/users.conf"
    cacheSpecification = "expireAfterWrite=30s"

Mixed Authentication

Mixed authentication allows collins to use multiple authentication backends simultaneously. An example use-case could be a scenario where real users live in LDAP but system users (for automation etc.) are defined using file authentication.

To use mixed authentication, simply specify a list of authentication backends as the type, and add configuration (as described above) for each of them within the authentication block. A simple example can be seen below.

The mixed authentication provider will try each of the methods in the order listed and use the first one that returns a successful match.

authentication {
  type = "file,ldap"

  file {
    userfile = "conf/users.conf"

  ldap {
    host = "localhost"
    groupAttribute = "uniqueMember"
    searchbase = "dc=example,dc=org"
    usersub = "ou=people"
    groupsub = "ou=groups"


Caching used by collins

Collins uses caching in several places, most notably to cache ldap responses, metrics, and to coordinate across multiple nodes for the firehose.

enabled Boolean Whether to enable the cache. Defaults to true
type String The type of cache to use. Valid options are in-memory for guava based cache or distributed for hazelcast based cache. in-memory is only valid for a standalone collins setup - if you have multiple collins servers, you should switch to using the distributed cache. Defaults to in-memory
cache {


Guava in-memory cache

Guava cache is only suitable for standalone setup. It is used to cache various information internally to collins.
enabled Boolean Whether to enable guava cache. Defaults to true. If you choose in-memory for your cache type, you must set this configuration to true as well.
specification String The global guava cache specification, as decribed in Defaults to "maximumSize=10000,expireAfterWrite=10s,recordStats"
guava {

  # used when deploying in standalone mode

Hazelcast distributed cache

Hazelcast is used for multi-node distributed cache. It was introduced to support the firehose feature.
enabled Boolean Whether to enable the hazelcast. Defaults to true. If you choose distributed for your cache type, you must set this configuration to true as well.
configFile String Path to the XML config file used by hazelcast. See for details. See for a working example.
members String List of IP addresses of existing hazelcast members. Make sure this includes at least one other collins IP, so the collins nodes can find each other.
hazelcast {

      # used when deploying in clustered mode


Hooking into lifecycle events in collins

Namespace callbacks

Collins supports a callback mechanism that allows you to trigger events when the configured callbacks are matched. A short example is probably most useful here.

callbacks.registry {
  hardwareProblem {
    on = "asset_update"
    when {
      previous.state = "!isMaintenance"
      current.state = "isMaintenance"
      current.states = [ HARDWARE_PROBLEM ]
    action {
      type = exec
      command = "/sbin/notify <tag>"

In the above configuration we have registered a callback named hardwareProblem. This callback will be triggered when an asset is moved to have a status of Maintenance with a state of HARDWARE_PROBLEM. When the callback is triggered we will execute a notify command that will email a systems administrator.

A callback is triggered by an event which is specified via the on config value. The value is then matched using a matcher which is specified via the when config value. If the matcher succeeds the specified action is taken.

enabled Boolean Defaults to true. Do not disable as some cache maintenance is done via callbacks.
class String The callback implementation to use. Defaults to collins.callbacks.CallbackManagerPlugin
registry Object A set of named objects describing various callback scenarios
callbacks {
  enabled = true
  class = "collins.callbacks.CallbackManagerPlugin"
  registry {

Registry Namespace callbacks.registry

The registry holds all your configured callbacks.

on String The condition to match on. Required. Valid values are below.
  • asset_create
  • asset_update
  • asset_delete
  • asset_log_create
  • asset_meta_value_create
  • asset_meta_value_delete
  • ipAddresses_create
  • ipAddresses_update
  • ipAddresses_delete
when Object Has previous state/states and current state/states as keys. These describe how to match whether the action should be triggered or not.
action Object Valid keys in the action are type (exec, collinscript) and a command


Collins feature flags

Namespace features

There are a number of parts of collins that you may wish to customize based on your environment. These are referred to in the reference configuration and elsewhere as feature flags. Available feature flags are documented below.

allowTagUpdates String A set of managed tags that can be updated for non-server assets
allowedServerUpdateStatuses Set[String] A set of statuses in which collins will accept LLDP/LSHW/CHASSIS_TAG updates for server assets. Maintenance is always a member of this set, even if not specified.
defaultLogType String Used for logs being created via the API when no type is specified
deleteIpmiOnDecommission Boolean Whether IPMI info should be deleted when an asset is decommissioned
deleteIpAddressOnDecommission Boolean Whether IP address information should be deleted on decommission
deleteMetaOnDecommission Boolean Whether attributes should be deleted on decommission
deleteSomeMetaOnRepurpose Set[String] If deleteMetaOnDecommission is false, this is a list of attributes to delete anyhow
encryptedTags Set[String] A list of attributes that should be encrypted. The ability to read and write these tags is controlled by the feature.canSeeEncryptedTags permission and the feature.canWriteEncryptedTags permission, respectively.
hideMeta Set[String] A list of tags to hide from display
ignoreDangerousCommands Set[String] A list of tags of assets that should ignore dangerous commands. You should populate this with tags of hosts that shouldn't be allowed to be powered off, reprovisioned, etc.
intakeSupported Boolean Whether the web based intake is supported or not
intakeIpmiOptional Boolean Whether the intake process may skip UID verification for assets without IPMI configuration.
keepSomeMetaOnRepurpose Set[String] If useWhitelistOnRepurpose is true, this is a list of attributes to keep
noLogAssets Set[String] A set of assets not to log any changes for
noLogPurges Set[String] A set of tags not to log when they change
searchResultColumns Set[String] A set of attributes to display on search results page. Defaults to [TAG, HOSTNAME, PRIMARY_ROLE, STATUS, CREATED, UPDATED]
sloppyStatus Boolean Should be true if you want collins to be usable.
sloppyTags Boolean Should probably be false. Allow every tag to be updated/overwritten/changed no matter what. Normally managed tags can only be updated during certain parts of the asset lifecycle.
useWhitelistOnRepurpose Boolean Determines which of deleteSomeMetaOnRepurpose or keepSomeMetaOnRepurpose is used. When true keepSomeMetaOnRepurpose is used.
features {
  allowTagUpdates = []

  allowedServerUpdateStatuses = []

  defaultLogType = Informational

  deleteIpmiOnDecommission = true

  deleteIpAddressOnDecommission = true

  deleteMetaOnDecommission = false

  deleteSomeMetaOnRepurpose = []

  encryptedTags = []

  ignoreDangerousCommands = []

  intakeSupported = true

  hideMeta = []

  noLogAssets = []

  noLogPurges = []

  sloppyStatus = true

  sloppyTags = false
features {
  allowedServerUpdateStatuses = [ PROVISIONED, PROVISIONING ]


  noLogAssets = [chatty_config, other_chatty_config]

  noLogPurges = [SYSTEM_PASSWORD]


Generic framing support for external tools

Overview Namespace views

In order to generically support external utilities to provide supplemental data for assets, a generic view system was implemented to allow operators to frame in arbitrary sites. This supplants the graphs plugin, and is used to provide easy access to tooling like Grafana, Nagios/Icinga, Prometheus, etc.

frames Map[String]Frame A map of frames to conditionally evaluate on any given asset. See examples.
enabled Boolean Defaults to false

Frames Namespace views.frames

enabled Boolean Enable this frame or not (default false)
title String Title to display on asset tab bar
style String HTML style tag to apply to the iframe. A good default is width: 100%;height: 1200px;
script String JS script to inject into the view to determine what the URL is to frame in, and how to decide if the tab should be marked as active or not. You can specify function isEnabled(jsonObj) and function getUrl(jsonObj). See examples.
views {
  enabled = true
  frames = {
		# frame in grafana
    graph {
      enabled = true
      title = "Grafana"
      style = "width: 100%;height: 1200px;"
      script = """
        // if the asset is a server, and has a hostname, enable this tab
        function isEnabled(inp) {
          var obj = JSON.parse(inp);
          return ((obj.ASSET.TYPE === "SERVER_NODE" || obj.ASSET.TYPE === "SWITCH") && ("HOSTNAME" in obj.ATTRIBS[0]));

        // all assets use the host-stats dashboard
        function getUrl(inp) {
          var obj = JSON.parse(inp);
          var dash = "host-stats"
          if (obj.ASSET.TYPE === "SWITCH") {
            dash = "switch-stats";
          return "" + dash + "?var-hostname=" + obj.ATTRIBS[0]["HOSTNAME"];

    # frame in nagios
    monitoring {
      enabled = true
      title = "Nagios"
      style = "width: 100%;height: 1200px;"
      script = """
        function isEnabled(inp) {
          var obj = JSON.parse(inp);
          return ((obj.ASSET.STATUS === "SERVER_NODE") && ("HOSTNAME" in obj.ATTRIBS[0]));
        function getUrl(inp) {
          var obj = JSON.parse(inp);
          return "" + obj.ATTRIBS[0]["HOSTNAME"];

The above example configures 2 tabs. One frames in Grafana, showing a different dashboard depending on if the asset is a Server or Switch. The other frames in Nagios for the Server asset. This can be trivially extended to support PagerDuty, Prometheus, Alertmanager, Sensu, and any other monitoring, graphing, or alerting system. All that is required is the ability to parameterize the URL to the system with some identifying attribute of the host (hostname, IP, asset tag, etc).


Display trending data

WARNING The graph namespace has been deprecated in favor of the more flexible and generic views namespace

Overview Namespace graph

Collins has an easy to implement plugin for graph and trending systems, which can be extended to support any framed in HTML content. The graph plugin will display embedded graphs or charts for assets, or just links to those systems (depending on the graph plugin). Once enabled, you will find an additional tab on the asset details screen named 'Graphs' with all the configured information available in it.

There is currently support for Fibr, the OpenTSDB visualization tool we built at Tumblr (currently closed source!), and ganglia. There are no plans to opensource Fibr at this time.

Asset trending

class String Class that implements the neccesary graph plugin. Defaults to collins.graphs.FibrGraphs
enabled Boolean Defaults to false.
cacheSpecification String The guava cache specification to use for metrics cache, as described in Defaults to "maximumSize=2000,expireAfterWrite=60s".

Fibr Graphs Namespace graph.FibrGraphs

At Tumblr, Fibr is our primary trending, graphing, and systems visualization tool. NOTE: this has been deprecated in favor of the more generalized framing system.

annotations Set[String] Annotations to display with graphs. Defaults to [deploy] (show deploy lines)
customMetrics Object Metrics to display based on asset information
defaultMetrics Set[String] Metrics to display for any asset. Defaults to SYS/LOAD, SYS/MEM, SYS/NET, SYS/PROC, SYS/IO-UTIL
timeframe Duration What timeframe to display for the metrics. Defaults to 1 hour (shows 1 hour of data).
url URL The Fibr web URL. Defaults to which is almost certainly not suitable outside of Tumblr.
graph.FibrGraphs {
  customMetrics = {
    mysqlHosts {
      selector = "PRIMARY_ROLE = \"DATABASE\""
      metrics = [
    memcacheHosts {
      selector = "POOL = MEMCACHE*"
      metrics = [

In the above example, hosts that have a PRIMARY_ROLE of DATABASE will display in addition to the default metrics, the MySQL specific metrics. Assets with a POOL matching MEMCACHE* will additionally display memcache specific metrics.

Ganglia Graphs Namespace graph.GangliaGraphs

Ganglia is a distributed metrics generation, collection, and visualization system. Collins integration relies on recent ad-hoc-views support.

Ganglia graph example

url String Location of the ganglia web ui.
timeRange String Default time range to display in the UI. Defaults to 1hr. Options available will depend on your ganglia setup.
hostSuffix String By default ganglia uses reverse dns to determine hostnames. That means a server with a hosname of foo might show up as If your reverse dns scheme has a consistent suffix, you can supply it here.
defaultGraphs Set[String] gweb veiws makes a distinction between "graphs" (or reports) and individual metrics. "graphs" are pre-defined and can combine multiple metrics whiel metrics just display a single value. Defaults to "load_all_report", "load_report", "mem_report", "cpu_report", "network_report", "packet_report".
defaultMetrics Set[String] Which individual metrics to display. Defaults to "disk_free", "disk_total".
graph {
  class = "collins.graphs.GangliaGraphs"
  enabled = true
  GangliaGraphs {
    url = "http://my-ganglia.domain.local/"

Ip address

IPAM Configuration

Namespace ipAddresses

The Collins IPAM configuration is pretty straight forward. You specify more or less a set of pools that are available for allocation. Once configured, assets can have addresses from any of these pools assigned to them.

defaultPoolName String If no pool name is specified, this is the one to use. Defaults to DEFAULT.
strict Boolean If true (default), refuse to create addresses in undefined pools. This is primarily useful to set to false if you don't control your IP space and you get addresses from a 3rd party.
pools Object A configured set of pools to allocate from
ipAddresses {
  pools {
    DEV {
      startAddress = ""
      network = ""
    ROW_1_1 {
      name = "ROW_1.1"
      startAddress = ""
      network = ""

The above configuration now allows you to allocate addresses out of two pools, a DEV pool and a ROW_1.1 pool. Notice that the ROW_1.1 pool had to specify a name. This is because an object name can not contain a period so must actually set the name manually as a string. The DEV pool did not have to do this. At Tumblr we have found pool names most useful for mapping to a particular VLAN and handling VLAN assignments using this information.

Pool Config

Each pool can take a name, gateway, startAddress and network.

name String The name of the pool. Defaults to the key of the pool.
network String The network to use for address allocation in CIDR notation. Defaults to
startAddress Option[String] The IP address to use to start allocating addresses at. Defaults to
gateway Option[String] If the gateway is not in the network range, specify it here


Collecting GPU information

Namespace gpu

GPU information can be stored in collins by adding GPU vendors to a whitelist. By default, any NVIDIA GPUs (PCI vendor = "NVIDIA Corporation") will be parsed from lshw XML data and added to collins. Additional vendors can be added by updating the configuration with their PCI vendor information.

supportedVendorStrings Set[String] PCI vendor strings to whitelist when parsing LSHW output. Defaults to ["NVIDIA Corporation"].
gpu {

  supportedVendorStrings = ["NVIDIA Corporation", "Advanced Micro Devices, Inc. [AMD/ATI]"]



Configuring your IPMI network and authentication setup

Namespace ipmi

The IPMI configuration is used to drive generation of IPMI addresses, usernames and passwords. This information is used for power management operations such as power cycling an asset or lighting up an asset during the intake process.

passwordLength Integer The length of the generated password. Defaults to 12.
randomUsername Boolean If false, uses the configured username. If true, generates a username based on the asset tag. Defaults to false.
username String Used when randomUsername is false. Defaults to root.
network String The network to use for address allocation in CIDR notation. Defaults to
startAddress Option[String] The IP address to use to start allocating addresses at. Defaults to
gateway Option[String] If the gateway is not in the network range, specify it here
ipmi {

  passwordLength = 12
  randomUsername = false
  username = "root"
  network = ""
  startAddress = ""



Configuration for importing LLDP data

Namespace lldp

Collins accepts XML output from lldpctl in order to store data about which switch an asset is connected to.

requireVlanName Boolean If true, throw an error if the LLDP data submitted contains any VLANs without a name attribute. Defaults to true.
requireVlanId Boolean If true, throw an error if the LLDP data submitted does not contain a VLAN with a vlan-id attribute. Defaults to false.
lldp {
     requireVlanName = true
     requireVlanId   = false


Monitoring Collins

Namespace metricsreporter

Monitoring Collins using Coda Hale's Metrics and Chris Burrough's metrics-reporter-config project is relatively straightforward. See Metrics' documentation for how to report specific metrics, and the metrics-reporter-config documentation for how the configuration file should be formatted.

enabled Boolean true to enable, false to disable. false by default.
configFile String The cofig file passed to metrics-reporter-config
metricsreporter {

  enabled= false
  configFile = "conf/metrics-reporter-config.yaml"



For when a single data center just isn't enough

Namespace multicollins

When Tumblr grew into a second datacenter, we didn't want to bother with having to setup multi-master replication across sites or do some kind of master site replication scheme. Instead what we decided is that we would make each Collins instance aware of each other, and allow them to query and interact with each other. Don't worry, no crazy zookeeper magic going on here.

If you have multiple independent collins installations setup, to make them aware of each other you have to do two things.

  1. Configure multi-collins
  2. Create an appropriate asset in each collins instance
enabled Boolean True to enable, false to disable. False by default.
instanceAssetType String Name of the asset type that represents a remote datacenter. Defaults to DATA_CENTER
locationAttribute String The tag on the instanceAssetType that provides collins with authentication information for the remote collins instance. This should be in URL format including credentials, e.g. http://username:password@remoteHost:remotePort. Don't worry, you can control who can see this with the features.canSeeEncryptedTags permission.
thisInstance String The asset tag that is for this specific instanceAssetType. For instance if the collins instance you are configuring is in EWR01, and the remote instance is in JFK01, you should set thisInstance to EWR01.
queryCacheTimeout Integer The amount of time (in seconds) to cache the query responses from remote collins instances. A value of 0 will result in values never expiring from the cache, which you probably do not want. Defaults to 30 seconds.
cacheEnabled Boolean Optionally enable or disable the multicollins cache. True, meaning enabled, by default.
multicollins {
  enabled = false
  instanceAssetType = DATA_CENTER
  locationAttribute = LOCATION
  thisInstance = NONE
  queryCacheTimeout = 30
  cacheEnabled = true

Once you have configured multicollins you will now be able to use the remoteLookup API parameter for asset finds. The Ruby client automatically handles fetching remote assets when needed. Additionally, you should find a 'Remote Search' checkbox in the search UI.

Node classifier

Find similar assets

Namespace nodeclassifier

A pretty common thing to want to be able to do is ask collins to tell you what other assets are like another. At Tumblr we use this for bulk provisioning, where I might want to be able to do something like say 'provisioning me 100 servers that are just like this one'. This is the problem that is solved with the node classifier. A node classifier describes a node class (e.g. web, queue, database, etc) and allows you to query for assets described by the classifier.

Much like with the MultiCollins configuration, we reuse assets to configure our node classifiers. In order to create a node classifier create a new asset (we prefer Configuration assets), and set tags on it that describe the traits of similar assets. For instance you might create a Configuration asset named web_classifier and set the NIC_SPEED to 10000000000, the MEMORY_SIZE_TOTAL to 34359738368, and the TOTAL_DISK_STORAGE to 500107862016. This configuration now describes a web class host with a 1Gb NIC, 32GB of RAM, and a 500GB disk. Be sure to set an IS_NODECLASS tag to true on the asset.

With the above configuration specified you can now use the find similar API endpoint. You also in the web UI should find links for finding similar assets.

assetType String The asset type for classifiers. Defaults to CONFIGURATION.
identifyingMetaTag String The tag that indicates that the asset is indeed a classifier. Defaults to IS_NODECLASS.
excludeMetaTags Set[String] A set of tags to exclude from finding similar assets. IS_NODECLASS and NOTE are included by default.
displayNameAttribute String What attribute to use for a friendly name for this classification, instead of the tag of the classification asset. Defaults to NAME.
sortKeys Set[String] Tags to be used for sorting. This is useful for instance for specifying that ROW, followed by RACK should be used for location based ordering of results. We use this at Tumblr to provide good distribution when provisioning.
nodeclassifier {
  identifyingMetaTag = IS_NODECLASS
  displayNameAttribute = NAME
  excludeMetaTags = [IS_NODECLASS, NOTE]
  sortKeys = []

Power configuration

Power requirements for assets

Namespace powerconfiguration


Depending on the environment, each asset may have a different power configuration. A power configuration specifies the number of power units required per asset, the number of power components required per power unit, and various requirements for each power component.


  • An asset may have an associated power configuration
  • A power configuration consists of one or more power units
  • A power unit consits of one or more power components (e.g. power strip, power outlet, etc)
  • A set of power units represents a complete power configuration

You can think of an asset power configuration as looking something like the following.

  • PowerConfiguration
    • Power Unit 1
      • Power Component 1
      • Power Component 2
      • Power Component n
    • Power Unit 2
    • Power Unit n


unitComponents Set[String] A set of components in a power configuration. Defaults to [PORT, OUTLET]
uniqueComponents Set[String] Components that must be unique for a single asset. Defaults to [PORT]
unitsRequired Integer Number of power units required for intake. Defaults to 1.
useAlphabeticNames Boolean Whether power units should use alphabetic names or not, defaults to true. The default configuration will result in POWER_PORT_A, POWER_OUTLET_A. If unitsRequired were 2 it would also include POWER_PORT_B and POWER_OUTLET_B. If useAlphabeticNames is false, instead of A/B it would be 1/2.

Note that this configuration heavily influences the intake process as well as what API parameters are available for asset updates. You can get a sample for what the asset looks like in the Asset Get JSON response.


It is possible to customize how the various components are labeled. The default bundled messages defines the following messages.

  • powerconfiguration.unit.port.label - Power Strip {0}
  • powerconfiguration.unit.outlet.label - Outlet {0}

Notice that the configured components each get a unit label. This allows you to customize how the intake screen looks.

Power managment

With great power, comes great responsibility

Namespace powermanagement

Assuming you would like the ability to perform power management operations on devices, such as power cycling them or powering them off, you will want to provide a power management configuration.

enabled Boolean Set to true to enable, defaults to false.
class String Defaults to "util.plugins.IpmiPowerManagement" which is probably suitable for your environment.
timeout Duration Defaults to 10 seconds. Do not increase as this will reduce concurrency on the system.
allowAssetTypes Set[String] Asset types that can be power managed. Defaults to [SERVER_NODE].
allowAssetsWithoutIpmi Boolean Allow power actions on assets that do not have IPMI information. Defaults to false.
disallowStatus Set[String] Set of status types where power management should not be enabled. Defaults to empty set.
disallowWhenAllocated Set[String] Set of power actions to disable when an asset is allocated. Defaults to powerOff.
commands Object Defines the commands to execute when a power action is called.
powermanagement {
  enabled = true
  command_template = "ipmitool -H <host> -U <username> -P <password> -I lan -L OPERATOR"

  allowAssetTypes = [SERVER_NODE]
  allowAssetsWithoutIpmi = false
  disallowStatus = []
  disallowWhenAllocated = [powerOff]

  commands {
    powerOff = ${powermanagement.command_template}" chassis power off"
    powerOn = ${powermanagement.command_template}" chassis power on"
    powerSoft = ${powermanagement.command_template}" chassis power soft"
    powerState = ${powermanagement.command_template}" chassis power status"
    rebootHard = ${powermanagement.command_template}" chassis power cycle"
    rebootSoft = ${powermanagement.command_template}" chassis power reset"
    identify = ${powermanagement.command_template}" chassis identify <interval>"
    verify = "ping -c 3 <host>"

Note A command must be specified for every power action.

powerOff Power off.
powerOn Power on.
powerSoft ACPI power off.
powerState Power status, should have 'on' in stdout if the device is on.
rebootHard Power cycle
rebootSoft ACPI reboot
identify Turn on the IPMI light for <interval> seconds
verify A command to run if the IPMI command fails. Should exit 0 if the host is reachable and exit 1 otherwise.
<tag> The tag of the asset being operated on
<username> The IPMI username associated with the asset
<password> The IPMI password associated with the asset
<host> The IPMI address associated with the asset
<interval> The interval to keep the light on for. 60 seconds.


Setup provisioning for your environment

Namespace provisioner

The provisioner plugin adds asset provisioning capabilities to collins which can be driven via the API or the web UI.

allowedStatus Set[String] Only machines that have a status listed here can be provisioned. Defaults to [Unallocated, Maintenance, Provisioning, Provisioned]
allowedType Set[String] Only assets with this type can be provisioned. Defaults to [SERVER_NODE, SERVER_CHASSIS]
enabled Boolean Defaults to false
cacheSpecification String The guava cache specification to use for the profiles file, as described in Defaults to "expireAfterWrite=30s".
checkCommand String A command that runs before the actual provison, should return 0 if provisioning can proceed or 1 otherwise. Defaults to /bin/true.
command String The command to run. Substitutions are of the form <foo> where foo is a supported substitution. Available substitutions are <tag>, <profile-id>, <notify>, <suffix> and <logfile>. Defaults to /bin/false
profiles String A YAML file containing all available provisioning profiles.
rate String How frequently users can provision hosts unless they are setup with the features.noRateLimit permission. Defaults to once every 10 seconds.
checkCommandTimeoutMs Duration Timeout for the execution of the checkCommand. Defaults to 40 seconds.
checkCommandTimeoutMs Duration Timeout for the execution of the command. Defaults to 40 seconds.
provisioner {

  allowedStatus = [Unallocated, Maintenance, Provisioning, Provisioned]


  enabled = false

  cacheSpecification = "expireAfterWrite=30s"

  checkCommand = "/bin/true"

  command = "/bin/false"

  profiles = "test/resources/profiles.yaml"

  rate = "1/10 seconds"

  # How long to wait for provisioning commands, checkCommand / command
  checkCommandtimeout = 40 seconds
  commandtimeout = 40 seconds

Provisioning Rates

Rates are expressed in terms of the number of successful requests per time period that a user can make. Specified as #/duration where # is an integer and duration is the duration. Examples:

  • 1/30 seconds - once per 30 seconds
  • 5/1 minute - five times per minute (same as 1/20 seconds)
  • 1/0 seconds - no rate limit
  • 0/0 seconds - disallowed

Substitutions in Commands

A few substitutions are available in the command and checkCommand

  • <tag> - the asset tag being provisioned
  • <profile-id> - the id of the profile being used for provisioning
  • <notify> - the contact information supplied by the user doing the provisioning
  • <suffix> - any suffix to append to the hostname as provided by the user (if allowed)
  • <logfile> - a logfile to use, defaults to /tmp/<tag>-<profile-id>.log

Provisioner Profile

The provisioner profiles file is what tells collins what types of assets can be provisioned. It is a fairly simple format and leaves most actual work up to the command and check command configured CLI tools. At Tumblr we have implemented a variety of simple CLI tools to handle a variety of provisioning scenarios. Our infrastructure automation framework is called visioner, some sample visions will be distributed so you can see how some of our provisioning works.

The keys in the reference file below (databasenode, devnode, webnode) are automatically set to be the NODECLASS tag of the asset in collins. This key is also reference above as <profile-id> and is how the provision process knows what configuration to use. A profile can have any number of key/values associated with them of any type, but has a few reserved key/value pairs.

    label: "Database Server"
    prefix: "db"
    primary_role: "DATABASE"
    requires_pool: false
    label: "Dev Machine"
    prefix: "dev"
    allow_suffix: true
    primary_role: "DEVEL"
    pool: "DEVEL"
      DEV_MODE: true
    label: "Hadoop Server"
    prefix: "hadoop"
    - storage_class
    - service_class
    label: "Web Server"
    prefix: "web"
    requires_secondary_role: true
label String Label to show in the UI. Should be unique.
prefix String The prefix to use in the hostname. This (semi-readable hostnames) helps with viewing error logs and such.
primary_role Option[String] The primary role to set on the asset in collins. The primary role drives automation and other tasks. If this value is set, the user can not select it.
secondary_role Option[String] The secondary role to set on the asset in collins. If this value is set, the end user can not select it.
pool Option[String] The pool to set on the asset in collins. If this value is set, the end user can not select it.
requires_primary_role Boolean Whether a primary role is required or not, defaults to true. If primary_role is not specified, the user must select one. If this is set to false, the user does not have to select one.
requires_secondary_role Boolean Whether a secondary role is required or not, defaults to false. If secondary_role is not specified and this value is true, the user must specify one. If this value is false, the user is not required to specify the value.
requires_pool Boolean Whether a pool is required or not, defaults to true. If pool is not specified and this value is true the user must specify one.
allow_suffix Boolean Whether a user may specify an additional suffix for the hostname, defaults to false. If this is set to true, and the user for instance specifies a value of "foo" and the prefix was set to "dev", the generated hostname will be something like "dev-foo".
contact Option[String] If specified, this value is set on the CONTACT attribute at provision time. Useful to map responsible parties to nodeclasses. If not specified, CONTACT is cleared on the asset.
contact_notes Option[String] Similar to contact, this field can be used to store some kind of information about the nodeclass, i.e. a link to a wiki page about the node profile, or services running on the machine. If not specified, this will be cleared at provision time.
attributes Map[String,Object] If specified, any key value pair here will be set as attributes on the asset at provision time. You can use any value types that YAML supports; the parsed representation will be turned into a string when setting the attribute. These attributes have a lower precedence than attributes set by collins, such as nodeclass and primary_role.
clear_attributes List[String] If specified, any attribute in this list will be cleared on provision. These attributes have a lower precedence than any attributes set in attributes, and wont clear any attributes set by collins itsself (nodeclass, primary_role, etc)
allowed_classes Option[List[String]] A list of asset classification tags that an asset must be classified as to allow provisioning. The list is ORed, so an asset matching any of the classifications will be provisionable. If omitted, there are no restrictions placed on what asset can be used to provision this profile (any classified/unclassified asset will provision). This is useful to restrict certain nodeclasses to certain hardware profiles, i.e. hadoopnode requires storage_class assets. See docs on asset classification for details on how this works.


Working with SoftLayer

Namespace softlayer

Although Collins was initially developed as Tumblr built out its own datacenter, we found it an excellent replacement for the SoftLayer web interface. The SoftLayer plugin provides support for cancelling servers, activating servers from the spare pool, and decorating tags that are SoftLayer specific.

Note The Collins system assumes that asset tags associated with SoftLayer assets are of the form sl-ID where ID is the SoftLayer ID.

enabled Boolean Defaults to false
username String Your SoftLayer username
password String Your SoftLayer API password
allowedCancelStatus Set[String] Assets with any of these status can be cancelled.
cancelRequestTimeoutMs Duration Timeout for canceling of a softlayer server. Defaults to 10 seconds.
activationRequestTimeoutMs Duration Timeout for activation of a softlayer server. Defaults to 10 seconds.
softlayer {
  enabled = true
  username = "jimbob123"
  password = "lkajsdlkjaas09823lkajdslkajsd"
  allowedCancelStatus = [Unallocated, Allocated, Maintenance]
  cancelRequestTimeout = 10 seconds
  activationRequestTimeout = 10 seconds

The above configuration would perform operations as user jimbob and allow you to cancel assets in the specified status.


Better search

Namespace solr

If the Solr plugin is enabled, Collins will use Solr to index assets. Since you can run Solr either embedded (bundled into Collins) or using an external instance, you should almost definitely keep this enabled.

Once the Solr plugin is enabled you will notice a few new features pop up in Collins. First, you can now query collins using the collins query language. This allows you to perform more advanced queries using ranges (how many assets have between 4 and 8 CPUs, give me assets with more than 2TB of usable storage) and much more sophisticated boolean operators. You can also sort results by any tag which is nice. You can find example queries in the collins query language section of the recipes guide.

More information about these configuration settings can be found on the official Solr documentation

embeddedSolrHome String Embedded only Must be a writable directory if useEmbeddedServer is true
enabled Boolean Defaults to true
externalUrl String External only If useEmbeddedServer is false, this must be a valid solr URL
reactToUpdates Boolean Keep index up to date as assets and values are modified. Defaults to true.
repopulateOnStartup Boolean Defaults to false. Useful in dev.
useEmbeddedServer Boolean Defaults to true
socketTimeout Int External only Solr socket timeout for HTTP client. Defaults to 1000ms.
connectionTimeout Int External only Solr HTTP connection timeout. Defaults to 5000ms.
maxTotalConnections Int External only Solr HTTP client tunable; defaults to 100.
defaultMaxConnectionsPerHost Int External only Solr HTTP client tunable; defaults to 100.
assetBatchUpdateWindowMs Int Window in ms that the Solr Updater will attempt to batch asset updates in. Defaults to 10ms. Increasing this can reduce load on Solr if intaking an asset is triggering multiple asset update tasks.

Tag decorators

Taking sexy back

Namespace tagdecorators

Tag decorators allow you to change how tags are displayed on asset. Creating links, formatting names, anything is fair game. The best way to illustrate is through an example.

tagdecorators {
  templates {
    search = "<a href=\"/resources?{name}={value}\">{value}</a>"
  decorators {
    APPLICATION_SHA.decorator    = "<a target=\"_blank\" href=\"{value}\">{value}</a>"
    CONTACT.decorator = "<a href=\"{value}\">{value}</a>"
    CONTACT_NOTES.decorator = "<a target=\"_blank\" href=\"{value}\">{value}</a>"
      decorator = ${}
      valueParser = "util.views.DelimiterParser"
      delimiter = "-"
      between = "-"
    NODECLASS.decorator = ${}
    POOL.decorator = ${}
    PRIMARY_ROLE.decorator = ${}
    SECONDARY_ROLE.decorator = ${}
    BUILD_CONTACT.decorator = "<a href=\"{value}\">{value}</a>"
      decorator="<a href=\"ssh://{value}\">>{value}</a>"
      between=", "
    HYPERVISOR.decorator  = "<a href=\"/resources?TAG={value}\">{value}</a>"
    AVAILABILITY_ZONE.decorator     = "<a href=\"/resources?IS_AVAILABILITY_ZONE=true&NAME={value}\">{value}</a>"

Thread pools

Thread Pool

Namespace akka

Configuring Collins akka pools is relatively straightforward. See the akka documentation for details. Collins makes use of a single akka dispatcher which can be configured in The dispatcher is shared for background processes and solr updates

Note Currently only on the development branch, scheduled for future release.

background-processor Configure for background processes, for instance ipmi invocations
solr_asset_updater Configure for solr updates on asset modification
solr_asset_log_updater Configure for updates on asset log modification
akka {
  actor {
    default-dispatcher = {
      fork-join-executor {
        parallelism-factor = 1.0

    deployment = {
      /background-processor = {
        dispatcher = default-dispatcher
        router = round-robin
        nr-of-instances = 128

      /solr_asset_updater = {
        dispatcher = default-dispatcher

      /solr_asset_log_updater = {
        dispatcher = default-dispatcher