| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
* metrics: add Prometheus metrics for concurrent upload tracking
Add Prometheus metrics to monitor concurrent upload activity for both
filer and S3 servers. This provides visibility into the upload limiting
feature added in the previous PR.
New Metrics:
- SeaweedFS_filer_in_flight_upload_bytes: Current bytes being uploaded to filer
- SeaweedFS_filer_in_flight_upload_count: Current number of uploads to filer
- SeaweedFS_s3_in_flight_upload_bytes: Current bytes being uploaded to S3
- SeaweedFS_s3_in_flight_upload_count: Current number of uploads to S3
The metrics are updated atomically whenever uploads start or complete,
providing real-time visibility into upload concurrency levels.
This helps operators:
- Monitor upload concurrency in real-time
- Set appropriate limits based on actual usage patterns
- Detect potential bottlenecks or capacity issues
- Track the effectiveness of upload limiting configuration
* grafana: add dashboard panels for concurrent upload metrics
Add 4 new panels to the Grafana dashboard to visualize the concurrent
upload metrics added in this PR:
Filer Section:
- Filer Concurrent Uploads: Shows current number of concurrent uploads
- Filer Concurrent Upload Bytes: Shows current bytes being uploaded
S3 Gateway Section:
- S3 Concurrent Uploads: Shows current number of concurrent uploads
- S3 Concurrent Upload Bytes: Shows current bytes being uploaded
These panels help operators monitor upload concurrency in real-time and
tune the upload limiting configuration based on actual usage patterns.
* more efficient
|
|
* Add bucket's traffic metrics
* Add bucket traffic to dashboards
* Fix bucket metrics help messages
* Fix variable names
|
|
Replaced `SeaweedFS_filer_` with `SeaweedFS_filerStore_` because the metric name was not found.
|
|
|
|
Corrected the dashboard to use the new metrics
The metric seems to have changed from SeaweedFS_filer_ to SeaweedFS_filerStore_. This commit replaces the ref in the dashboard with the new series name.
|
|
|
|
|
|
|
|
comes from https://github.com/chrislusf/seaweedfs/pull/2511
|
|
In the current dashboard there is no way to know the status of
volume servers, as prometheus only accepts push information, if
any server volume goes offline, the last information sent will
be read continuously by grafana.
Added a graph that calculates the delta between the information
of the last push timestamp with the timestamp of 15 seconds ago,
if the timestamp has not been changed, server state will be 0,
if volume server sent any information the server state will be 1.
-This graph will have to be modified if the time interval of pushs
is modified, a suggestion for improvement is to send the interval
in seconds along with the metrics.
-Only local tests were run, maybe if the latency between pushgateway
and volume server is too high, the delta calculation could be
affected.
|
|
|
|
|
|
|
|
|
|
|
|
|
