RTFS Configuration

This reference lists all configuration properties for the Repository Federation Service (RTFS). These are set in application.properties or via Helm chart values under persistence.federationProperties.

Overriding Properties via system.yaml or the Platform Helm Chart

You can override RTFS application.properties settings without editing files inside the RTFS container. On native installations, Docker, and other deployments that use the Artifactory system.yaml, add properties under rtfs.persistence.federationProperties as key-value pairs:

rtfs:
  enabled: true
  persistence:
    federationProperties:
      rtfs.auto.healing.enabled: "true"
      rtfs.full.sync.batch.size: "2000"
      rtfs.file.list.published.ttl.min: "30"

Helm installations:

When RTFS is enabled through the Artifactory or platform Helm chart and system.yaml overlays, set the key-value map under rtfs.persistence.federationProperties in your chart values (for example, custom-values.yaml), then run helm upgrade.
When you deploy the standalone Artifactory Federation Service Helm chart, set the same keys under persistence.federationProperties, not under rtfs.. For more information, see Helm Deployment Configuration.

For RTFS enablement and database settings, see RTFS Database Requirements.

Apply changes: Restart Artifactory after updating system.yaml or Helm values. The RTFS service restarts with the platform.

Scope: Any valid RTFS application.properties key can be set through federationProperties. Common examples include auto-healing (rtfs.auto.healing.enabled), full sync batch size (rtfs.full.sync.batch.size), full sync file list TTL (rtfs.file.list.published.ttl.min), and worker lag thresholds (rtfs.workers.lag.threshold.sec). For advanced tuning beyond these examples, contact JFrog Support.

Full Sync

The following properties configure Full Sync behavior.

Property	Default	Description
`rtfs.full.sync.batch.size`	2000	Artifacts per full sync batch.
`rtfs.full.sync.job.run.members.batch.size`	50	Members to process per full sync job run.
`rtfs.full.sync.job.interval.sec`	120	Interval between full sync job runs (seconds).
`rtfs.full.sync.job.initial.delay.min`	1	Initial delay before first full sync job (minutes).
`rtfs.full.sync.grace.period.min`	60	Minimum time between full syncs for the same member (minutes).
`rtfs.full.sync.execution.time.threshold.min`	30	Threshold for logging long-running full syncs (minutes).
`rtfs.full.sync.batch.retry.quite.period.seconds`	5	Quiet period between batch retries (seconds).
`rtfs.full.sync.batch.retry.count`	5	Maximum batch retry count.
`rtfs.full.sync.heartbeat.stale.min`	30	Time before a full sync heartbeat is considered stale (minutes).
`rtfs.full.sync.cleanup.cron`	`0 0/10 * * * ? *`	Cron expression for full sync cleanup job.
`rtfs.dry.full.sync.batch.size`	2000	Batch size for dry full sync.
`federated.repo.full.sync.backoff.time.min`	0	Backoff time between full syncs (minutes).

Auto-Healing

The following properties configure auto-healing for members in error states.

Property	Default	Description
`rtfs.auto.healing.enabled`	true	Enable automatic recovery for members in error states.
`rtfs.auto.healing.job.interval.sec`	20	Auto-healing check interval (seconds).
`rtfs.auto.healing.job.initial.delay.min`	1	Initial delay before first auto-healing check (minutes).

Event Stream and Workers

The following properties configure event stream processing and worker batches.

Property	Default	Description
`rtfs.stream.allocation.size`	50	Stream allocation size for event processing.
`rtfs.worker.batch.size`	100	Events per worker batch (STANDARD priority).
`rtfs.worker.max.events.processed.in.batch`	200	Maximum events processed in a single batch.
`rtfs.worker.max.events.processed.in.batch.for.high.priority`	400	Maximum events for HIGH or URGENT priority batches.
`rtfs.worker.max.retry.count`	2	Maximum retry count per worker.
`rtfs.worker.min.return.backoff.sec`	2	Minimum backoff between retries (seconds).
`rtfs.worker.batch.processing.time.margin.sec`	5	Time margin for batch processing (seconds).
`rtfs.skip.delete.events.dedup`	false	Skip deduplication of delete events.
`rtfs.events.processor.request.timeout.sec`	30	Timeout for event processing requests (seconds).
`rtfs.events.cleanup.cron`	`0 0 1 * * ?`	Cron for event cleanup job.
`rtfs.events.cleanup.retention.time.Days`	3	Event retention time (days).

Stream Recovery and Lag

Stream recovery automatically detects and restarts stalled event processing streams. Stream lag monitoring tracks the delay between event generation and processing, identifying repositories that are falling behind.

When a worker stream stops processing events (for example, due to a transient database issue), the stream recovery job detects the stall and restarts the stream. The lag collector measures the time difference between the oldest unprocessed event and the current time, reporting this as the jfrtfs_worker_lag metric.

Property	Default	Description
`rtfs.stream.recovery.job.cron`	`/10 * ? * * *`	Cron for stream recovery job. Runs every 10 seconds to detect and restart stalled streams.
`rtfs.stream.lag.collector.job.cron`	`0 /2 ? * *`	Cron for stream lag collection. Runs every 2 minutes to measure processing delay.
`rtfs.stream.lag.stale.grace.period.sec`	120	Grace period before a stream's lag data is considered stale (seconds). After this period without updates, the lag data may be discarded.
`rtfs.workers.lag.threshold.sec`	20	Worker lag threshold (seconds). When lag exceeds this value, the member may transition to a DELAYED status in the dashboard.
`rtfs.workers.lag.top.repos.limit`	3	Number of top lagging repos to report in log entries, helping identify which repositories are most behind.
`rtfs.worker.lag.log.period.minutes`	30	Interval between lag summary log entries (minutes).
`rtfs.worker.lag.log.threshold.minutes`	1	Minimum lag before a repository appears in lag log entries (minutes).

HTTP Client

The following properties configure the RTFS HTTP client.

Property	Default	Description
`rtfs.reactive.http.client.connect.timeout.sec`	120	Connection timeout (seconds).
`rtfs.reactive.http.client.response.timeout.sec`	120	Response timeout (seconds).
`rtfs.reactive.http.client.read.timeout.sec`	120	Read timeout (seconds).
`rtfs.reactive.http.client.write.timeout.sec`	120	Write timeout (seconds).
`rtfs.reactive.http.client.max.in.memory.size`	5MB	Maximum in-memory buffer size.
`rtfs.reactive.http.client.max.connections`	50	Maximum HTTP connections.
`rtfs.reactive.http.client.max.idle.time.sec`	20	Maximum idle time (seconds).
`rtfs.reactive.http.client.max.life.time.sec`	30	Maximum connection lifetime (seconds).
`rtfs.reactive.http.client.evict.in.background.sec`	30	Background eviction interval (seconds).
`rtfs.reactive.http.client.retry.count`	5	HTTP retry count.
`rtfs.reactive.http.client.retry.backoff.ms`	100	Retry backoff (milliseconds).
`rtfs.reactive.http.client.retry.error.patterns`	Connection prematurely closed, Connection reset, Connection refused, Connection timed out	Error patterns that trigger retries.
`rtfs.http.client.connection.timeout.ms`	5000	HTTP client connection timeout (milliseconds).
`rtfs.http.client.socket.timeout.sec`	300	HTTP socket timeout (seconds).

Binary Tasks

The following properties configure binary task processing.

Property	Default	Description
`rtfs.binary.tasks.enabled.status.cache.expiry.seconds`	300	Cache expiry for binary tasks enabled status.
`rtfs.binary.tasks.legacy.cache.expiry.seconds`	300	Cache expiry for legacy binary tasks.
`rtfs.binary.tasks.inbound.insert.batch.size`	500	Batch size for inbound binary task inserts.
`rtfs.binary.tasks.handler.initial.delay.duration`	PT1M	Initial delay for binary task handler.
`rtfs.binary.tasks.handler.run.duration`	PT1S	Run interval for binary task handler.
`rtfs.binary.tasks.handler.tenants.thread.pool.size`	100	Thread pool size for tenant binary tasks.
`rtfs.binary.tasks.handler.fetch.limit.per.tenant`	500	Max binary tasks fetched per tenant.
`rtfs.binary.tasks.tenant.penalty.enabled`	true	Enable tenant penalty for binary tasks.
`rtfs.binary.tasks.tenant.penalty.duration`	PT1S	Penalty duration.
`rtfs.binary.tasks.tenant.penalty.recovery.duration`	PT1S	Penalty recovery duration.
`rtfs.binary.tasks.max.retry.per.task`	10	Maximum retries per binary task.
`rtfs.binaries.tasks.metrics.pending.lag.threshold.duration`	PT5M	Pending lag threshold for metrics.
`rtfs.binaries.tasks.metrics.pending.lag.collection.duration`	PT1M	Pending lag collection interval.
`rtfs.binaries.tasks.metrics.executing.lag.threshold.duration`	PT5M	Executing lag threshold for metrics.
`rtfs.binaries.tasks.metrics.executing.lag.collection.duration`	PT1M	Executing lag collection interval.

File List Tasks

The following properties configure file list task processing.

Property	Default	Description
`rtfs.file.list.tasks.job.cron`	`/30 * ? * * *`	Cron for file list tasks job.
`rtfs.file.list.tasks.allowed`	30	Maximum concurrent file list tasks.
`rtfs.file.list.heartbeat.expired.minutes`	15	Heartbeat expiry for file list tasks.
`rtfs.file.list.polling.limit.time.hours`	24	Maximum polling time for file lists.
`rtfs.file.list.polling.wait.time.seconds`	30	Wait time between file list polls.
`rtfs.file.list.size.metrics.threshold`	1000000	File list size threshold for metrics.
`rtfs.file.list.process.rate.threshold.artifact.per.sec`	5	Processing rate threshold.
`rtfs.file.list.sort.rate.threshold.artifact.per.sec`	10	Sort rate threshold.
`rtfs.sync.in.memory.sort.enabled`	true	Enable in-memory file list sorting. Required when performing Full Sync from a file.
`rtfs.sync.amount.of.entries.in.file`	400000	Entries per file list file.
`file.list.tasks.thread.pool.size`	30	Thread pool for file list tasks.
`rtfs.file.list.published.ttl.min`	30	Time-to-live for a published Full Sync file list, in minutes. When a new Full Sync from a file is requested for a member within this window, RTFS reuses the existing published file list instead of regenerating it. Set to `0` to disable reuse and always regenerate the file list. See Trigger a Full Sync.

Heartbeat and Mirror

The following properties configure heartbeat and mirror behavior.

Property	Default	Description
`rtfs.mirror.heartbeat.cron`	`/20 * ? * * *`	Cron for mirror heartbeat job.
`rtfs.mirror.heartbeat.stale.sec`	600	Stale heartbeat threshold (seconds).
`rtfs.mirror.heartbeat.inactive.threshold.hours`	24	Inactive member threshold (hours).
`rtfs.enable.disable.check.expiry.time.min`	15	Enable or disable check expiry (minutes).

Priority

The following properties configure repository priority limits.

Property	Default	Description
`rtfs.max.high.priority.repos.per.member`	10	Maximum HIGH priority repos per member.
`rtfs.max.high.priority.repos.percentage`	0.2	Maximum HIGH priority repos as percentage.

Thread Pools

The following properties configure RTFS thread pool sizes.

Property	Default	Description
`rtfs.jobs.thread.pool.size`	100	General jobs thread pool.
`rtfs.heartbeat.thread.pool.size`	100	Heartbeat thread pool.
`rtfs.tenant.thread.pool.size`	100	Tenant operations thread pool.
`rtfs.tenant.heartbeat.thread.pool.size`	200	Tenant heartbeat thread pool.
`rtfs.full.sync.thread.pool.size`	30	Full sync thread pool.
`rtfs.request.tunneling.thread.pool.size`	30	Request tunneling thread pool.
`rtfs.support.bundle.thread.pool.size`	5	Support bundle thread pool.
`rtfs.tenant.config.sync.job.thread.pool.size`	100	Tenant config sync thread pool.
`rtfs.repo.event.thread.pool.size`	100	Repo event thread pool.

gRPC

The following property configures the gRPC payload size.

Property	Default	Description
`rtfs.grpc.server.max.payload.size`	4194304 (4MB)	Maximum gRPC payload size.

Caching

The following properties configure RTFS caching behavior.

Property	Default	Description
`rtfs.jpd.discovery.cache.time.min`	15	JPD discovery cache time (minutes).
`rtfs.metric.cache.expiry.min`	2	Metric cache expiry (minutes).
`rtfs.traffic.extended.info.cache.time.min`	120	Traffic extended info cache time.
`rtfs.tenant.queue.exists.cache`	5	Tenant queue existence cache time.
`rtfs.repo.mgmt.cache.expiration.time.min`	40	Repo management cache expiry (minutes).
`rtfs.accessClient.remoteTokensCache.expirationSeconds`	600	Remote tokens cache expiry (seconds).

Scheduled Jobs

The following properties configure RTFS scheduled jobs.

Property	Default	Description
`rtfs.call.home.collect.metrics.job.cron`	`0 0 3 * * ?`	Call home metrics collection.
`rtfs.collect.metrics.job.cron`	`/59 * ? * * *`	General metrics collection.
`rtfs.orphan.cleanup.cron`	`0 0 2 * * ?`	Orphan cleanup job.
`rtfs.partition.job.interval.cron`	`0 0/30 * ? * *`	Partition management job.
`rtfs.sub.partition.pk.monitoring.cron`	`0 0 0 ? * * *`	Sub-partition PK monitoring.
`rtfs.repo.config.sync.job.interval.cron`	`0 0 * * * ?`	Repo config sync job.
`rtfs.repo.config.integrity.job.interval.cron`	`0 30 * * * ?`	Repo config integrity check.
`rtfs.readiness.job.cron`	`/15 * ? * * *`	Readiness probe job.
`rtfs.export.metrics.job.interval.sec`	60	Metrics export interval (seconds).
`rtfs.export.metrics.job.initial.delay.min`	1	Metrics export initial delay (minutes).

Miscellaneous

The following properties configure other RTFS behavior.

Property	Default	Description
`rtfs.validate.configuration.presence.enabled`	true	Validate configuration presence on startup.
`rtfs.repo.config.v2.jobs.enabled`	true	Enable v2 repo config jobs.
`rtfs.replication.log.interval.ms`	1000	Replication logging interval.
`rtfs.chunk.of.sub.partition.size`	7000000	Sub-partition chunk size.
`rtfs.threshold.percentage.partition.creation.size`	50	Partition creation threshold percentage.
`rtfs.db.lock.native.pool.size`	3	Database lock pool size.
`rtfs.local.jpd.cache.fetch.url.max.attempts`	3	Max URL fetch attempts.
`rtfs.local.jpd.cache.fetch.url.backoff.interval.ms`	300	URL fetch backoff interval.
`rtfs.request.tunneling.allowlist`	""	Request tunneling allowlist.
`bi.directional.connectivity.enable`	false	Bidirectional connectivity check.
`rtfs.traffic.rollover.thread.enabled`	true	Enable traffic rollover thread.
`rtfs.tenant.move.export.lag.offset.grace.period.min`	10	Tenant move export lag grace period.
`rtfs.exchange.token.limit.scope.enabled`	true	Limit token exchange scope.
`rtfs.batch.mirror.forward.client.ip.header.enabled`	false	Forward client IP header in batch mirrors.
`rtfs.bridge.url.enabled`	false	Enable bridge URL.
`rtfs.access.enable.platform.config.cache`	true	Enable platform config cache.
`rtfs.access.bridge.retry.backoff.initial.interval.ms`	20	Access bridge retry initial interval.
`rtfs.access.bridge.retry.backoff.max.interval.ms`	1000	Access bridge retry max interval.
`rtfs.access.bridge.retry.backoff.multiplier`	2.0	Access bridge retry multiplier.
`rtfs.artifactory.bridge.unavailable.retry.max.attempts`	4	Artifactory bridge retry max attempts.
`rtfs.artifactory.bridge.unavailable.retry.backoff.sec`	60	Artifactory bridge retry backoff.