-
Bug
-
Resolution: Fixed
-
Major
-
-
2.379
Agents re-connect logic is trying to connect hundreds of times per second when using websocket connection type.
When there is a persistent problem right after the connection is established (ie. agent running on Java 8 and Jenkins on 11), the re-try logic is just generating unnecessary load on the controller.
Trying to re-connect every 10 seconds (like the inbound tcp re-connection logic does) should be enough.
- links to
[JENKINS-70105] Slow down re-connects when using websockets
Assignee | Original: Jeff Thompson [ jthompson ] | New: Antonio Muñiz [ amuniz ] |
Status | Original: Open [ 1 ] | New: In Progress [ 3 ] |
Remote Link | New: This issue links to "PR (Web Link)" [ 28355 ] |
Description |
Original:
Agents re-connect logic is trying to connect hundreds of times per second when using websocket connection type.
When there is a persistent problem right after the connection is established (ie. agent running on Java 8 and Jenkins on 11), the re-try logic is just generating unnecessary load on the controller. Trying to re-connect every 10 seconds ([https://github.com/jenkinsci/remoting/blob/7e9b0dc0846678cdf08717710ba27b0e6e443911/src/main/java/hudson/remoting/Engine.java#L856|http://example.com]) should be enough. |
New:
Agents re-connect logic is trying to connect hundreds of times per second when using websocket connection type.
When there is a persistent problem right after the connection is established (ie. agent running on Java 8 and Jenkins on 11), the re-try logic is just generating unnecessary load on the controller. Trying to re-connect every 10 seconds ([https://github.com/jenkinsci/remoting/blob/7e9b0dc0846678cdf08717710ba27b0e6e443911/src/main/java/hudson/remoting/Engine.java#L856|like the inbound tcp re-connection logic does]) should be enough. |
Description |
Original:
Agents re-connect logic is trying to connect hundreds of times per second when using websocket connection type.
When there is a persistent problem right after the connection is established (ie. agent running on Java 8 and Jenkins on 11), the re-try logic is just generating unnecessary load on the controller. Trying to re-connect every 10 seconds ([https://github.com/jenkinsci/remoting/blob/7e9b0dc0846678cdf08717710ba27b0e6e443911/src/main/java/hudson/remoting/Engine.java#L856|like the inbound tcp re-connection logic does]) should be enough. |
New:
Agents re-connect logic is trying to connect hundreds of times per second when using websocket connection type.
When there is a persistent problem right after the connection is established (ie. agent running on Java 8 and Jenkins on 11), the re-try logic is just generating unnecessary load on the controller. Trying to re-connect every 10 seconds ([like the inbound tcp re-connection logic does|https://github.com/jenkinsci/remoting/blob/7e9b0dc0846678cdf08717710ba27b0e6e443911/src/main/java/hudson/remoting/Engine.java#L856]) should be enough. |
Status | Original: In Progress [ 3 ] | New: In Review [ 10005 ] |
Issue Type | Original: Improvement [ 4 ] | New: Bug [ 1 ] |
Released As | New: 2.379 | |
Resolution | New: Fixed [ 1 ] | |
Status | Original: In Review [ 10005 ] | New: Fixed but Unreleased [ 10203 ] |
Labels | New: lts-candidate |