-* `enabled`: If set to `true`, failed federation jobs will be retried
-* `max_jobs`: The maximum amount of parallel federation jobs running at the same time.
-* `initial_timeout`: The initial timeout in seconds
-* `max_retries`: The maximum number of times a federation job is retried
+### Note on running with PostgreSQL in silent mode
+
+If you are running PostgreSQL in [`silent_mode`](https://postgresqlco.nf/en/doc/param/silent_mode?version=9.1), it's advised to set [`log_destination`](https://postgresqlco.nf/en/doc/param/log_destination?version=9.1) to `syslog`,
+otherwise `postmaster.log` file may grow because of "you don't own a lock of type ShareLock" warnings (see https://github.com/sorentwo/oban/issues/52).
+
+## :workers
+
+Includes custom worker options not interpretable directly by `Oban`.
+
+* `retries` — keyword lists where keys are `Oban` queues (see above) and values are numbers of max attempts for failed jobs.
+
+Example:
+
+```elixir
+config :pleroma, :workers,
+ retries: [
+ federator_incoming: 5,
+ federator_outgoing: 5
+ ]
+```
+
+### Migrating `Pleroma.Web.Federator.RetryQueue` settings
+
+* `max_retries` is replaced with `config :pleroma, :workers, retries: [federator_outgoing: 5]`
+* `enabled: false` corresponds to `config :pleroma, :workers, retries: [federator_outgoing: 1]`
+* deprecated options: `max_jobs`, `initial_timeout`