You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
After running edge for a while I spotted that it was suddenly sleeping for 62 seconds, and then 64 seconds. Double checking I see that we add jitter to previous interval instead of resetting back to initial configured duration before adding jitter.
Steps to reproduce the bug
Start edge with RUST_LOG=warn,unleash_edge=trace
Observe
2024-02-29T12:09:33.846888Z TRACE unleash_edge::http::background_send_metrics: Done posting traces. Sleeping for 62 seconds and then going again
2024-02-29T12:09:33.846888Z TRACE unleash_edge::http::background_send_metrics: Done posting traces. Sleeping for 70 seconds and then going again
?
Profit?
Expected behavior
Interval should never exceed initial configured interval + max_jitter unless we're erroring. Since all posts succeeded, we probably shouldn't add jitter at all. the normal interval due to random startup times should be enough
Logs, error output, etc.
2024-02-29T12:09:33.846888Z TRACE unleash_edge::http::background_send_metrics: Done posting traces. Sleeping for 70 seconds and then going again
Screenshots
No response
Additional context
No response
Unleash version
Edge 17.1.0
Subscription type
None
Hosting type
None
SDK information (language and version)
No response
The text was updated successfully, but these errors were encountered:
Previously we used the previous iteration of the loop's interval to
decide next interval. This PR changes to always use the initial
send_interval as what to base the interval calculation of
fixes: #436
Previously we used the previous iteration of the loop's interval to
decide next interval. This PR changes to always use the initial
send_interval as what to base the interval calculation of
fixes: #436
Describe the bug
After running edge for a while I spotted that it was suddenly sleeping for 62 seconds, and then 64 seconds. Double checking I see that we add jitter to previous interval instead of resetting back to initial configured duration before adding jitter.
Steps to reproduce the bug
RUST_LOG=warn,unleash_edge=trace
Expected behavior
Interval should never exceed initial configured interval + max_jitter unless we're erroring. Since all posts succeeded, we probably shouldn't add jitter at all. the normal interval due to random startup times should be enough
Logs, error output, etc.
Screenshots
No response
Additional context
No response
Unleash version
Edge 17.1.0
Subscription type
None
Hosting type
None
SDK information (language and version)
No response
The text was updated successfully, but these errors were encountered: