Overlord Process
Configuration
For Apache Druid Overlord Process Configuration, see Overlord Configuration.
For basic tuning guidance for the Overlord process, see Basic cluster tuning.
HTTP endpoints
For a list of API endpoints supported by the Overlord, please see the API reference.
Overview
The Overlord process is responsible for accepting tasks, coordinating task distribution, creating locks around tasks, and returning statuses to callers. Overlord can be configured to run in one of two modes - local or remote (local being default). In local mode Overlord is also responsible for creating Peons for executing tasks. When running the Overlord in local mode, all MiddleManager and Peon configurations must be provided as well. Local mode is typically used for simple workflows. In remote mode, the Overlord and MiddleManager are run in separate processes and you can run each on a different server. This mode is recommended if you intend to use the indexing service as the single endpoint for all Druid indexing.
Blacklisted workers
If a MiddleManager has task failures above a threshold, the Overlord will blacklist these MiddleManagers. No more than 20% of the MiddleManagers can be blacklisted. Blacklisted MiddleManagers will be periodically whitelisted.
The following variables can be used to set the threshold and blacklist timeouts.
druid.indexer.runner.maxRetriesBeforeBlacklist
druid.indexer.runner.workerBlackListBackoffTime
druid.indexer.runner.workerBlackListCleanupPeriod
druid.indexer.runner.maxPercentageBlacklistWorkers
Autoscaling
The Autoscaling mechanisms currently in place are tightly coupled with our deployment infrastructure but the framework should be in place for other implementations. We are highly open to new implementations or extensions of the existing mechanisms. In our own deployments, MiddleManager processes are Amazon AWS EC2 nodes and they are provisioned to register themselves in a galaxy environment.
If autoscaling is enabled, new MiddleManagers may be added when a task has been in pending state for too long. MiddleManagers may be terminated if they have not run any tasks for a period of time.