Allow config of the backoff algorithm for the federation client. (#15754)

Adds three new configuration variables:

* destination_min_retry_interval is identical to before (10mn).
* destination_retry_multiplier is now 2 instead of 5, the maximum value will
  be reached slower.
* destination_max_retry_interval is one day instead of (essentially) infinity.

Capping this will cause destinations to continue to be retried sometimes instead
of being lost forever. The previous value was 2 ^ 62 milliseconds.

1 files changed, 11 insertions, 0 deletions

diff --git a/docs/usage/configuration/config_documentation.md b/docs/usage/configuration/config_documentation.md
index 4e6fcd085a..c32608da2b 100644
--- a/docs/usage/configuration/config_documentation.md
+++ b/docs/usage/configuration/config_documentation.md
@@ -1242,6 +1242,14 @@ like sending a federation transaction.
 * `max_short_retries`: maximum number of retries for the short retry algo. Default to 3 attempts.
 * `max_long_retries`: maximum number of retries for the long retry algo. Default to 10 attempts.
 
+The following options control the retry logic when communicating with a specific homeserver destination.
+Unlike the previous configuration options, these values apply across all requests
+for a given destination and the state of the backoff is stored in the database.
+
+* `destination_min_retry_interval`: the initial backoff, after the first request fails. Defaults to 10m.
+* `destination_retry_multiplier`: how much we multiply the backoff by after each subsequent fail. Defaults to 2.
+* `destination_max_retry_interval`: a cap on the backoff. Defaults to a week.
+
 Example configuration:
 ```yaml
 federation:
@@ -1250,6 +1258,9 @@ federation:
   max_long_retry_delay: 100s
   max_short_retries: 5
   max_long_retries: 20
+  destination_min_retry_interval: 30s
+  destination_retry_multiplier: 5
+  destination_max_retry_interval: 12h
 ```
 ---
 ## Caching