Branch 3.7.0

Branch 3.7.0

• set prometheus version to 2.25.2 #1333
• start-grafana.sh: set grafana to version 7.4.5 #1298
• Update the alertmanager plugin to 1.0 #1288
• Add the CQL errors to the dashboards #1276
• Add more info to client table enhancement scylla 4.4 #1259
• Provide some indication of hinted handoff being accumulated / being sent #1258
• Secondary Indexs / Materialized Views maybe be built in the background #1257
• Add manager task progress indication to the overview dashboard #1250
• prometheus.rules.yml add node restart alert#1248
• Provide indication of coordinator / replica errors per node #1229
• Print more info about backups #1180
• Switch the alertmanager to the new table panels #1071
• Visually present error/no error on the node table Need Info #1035
• Need Panels for Scylla HWLB #907
• Make the overview dashboard the default #677

Advisor

• The balance queries should be per DC not per cluster #1312
• Add report an issue from the advisor #1279
• Open an issue should be only for OS users #1311
• Warn users on non paged queries passing the soft limit / reaching hard limit and being aborted #1246
• Warn user that reverse reads are crossing the soft limit /reaching the hard limit and are aborted #1245
• Alert: Add an alert that scylla was restarted - it can be in low priority since the source maybe be an ops operation enhancement #1243
• OOM Killer kills Scylla / Scylla-JMX - validate whats running / memory setup advisor #1242
• Add info to each of the balance signs #1166

Bug Fixes

• loki container breaks -A option in start-all.sh #1326
• inconsistency use of legends in DC panel #1290
• Disk space alerts not working due to wrong metrics used #1282
• When using private network loki and alertmanager does not work #1252

Branch 3.6.3

Branch 3.6.3

• Dashboard for Scylla Manager 2.3

Bug fixes

• loki container breaks -A option in start-all.sh #1326

Branch 3.6.2

Bug fixes

• Timeouts and latencies per shards panels are missing #1294
• Non Token Aware queries for counters - A work around for #804

Branch 3.6.1

New Dashboards

• Adding dashboard for Scylla Open-source 4.4 #1271

Bug fixes

• Write latency and write count should not include hints/streaming scheduling group #1265
• Update all advisor / cql dashboard queries taking into account only the user gerenated queries and not internal ones #1263

Those bug fixes are relevant for Scylla Open-source 4.2, 4.3, 4.4 users and for Sclla enterprise 2020.1 users

Branch 3.6

New in Branch 3.6

• Use loki as data source #1147
• Adding the Advisor section #1162
• Add Scylla 4.3 dashboards #1144
• Migrate the nodes table to use the new table format #1097
• Grafana: Use UTC by default #1065
• Upgrade to Grafana 7.3.5 #1061
• Collapsible panels #973
• Support dynamic intervals #957
• New LWT metrics for the dasboard #936
• Add Loki log collection service to the monitoring stack #617
• Use docker-compose as an optional replacement for start-all.sh #273
• Easy way to capture the entire dashboard, or even all the dashboards, in one click #248
• Allow supply more Prometheus targets file from the command line #1197

Bug Fixes

• Passing --no-loki got illegal option: #1152

Branch 3.5.3

New in 3.5.3

Dashboard for Scylla Open source 4.3

Branch 3.5.2

Bug Fixes

• prometheus.yml.template: remove the second manager job and leave the migration to the users #1150
• start-grafana.sh looks for the docker IP which breaks on Podman #1145
• Alternator dashboard node-table should use the names of the new dashboards #1134

For user migrating to Scylla-Manager 2.2

Note that Scylla-Manager changed its port and you need to update scylla manager server file.

Branch 3.5.1

• sstable reads units are wrong #1125
• Links on the "Nodes" table on "Overview" don't go to the correct place #1122
• Hack that added support for the old and new ports doesn't work with the Agent graphs on the Manager dashboard #1120
• The -N flag is ignored in start-all.sh #1116
• Remove the avg line from the multi-graph panels in the overview #1115
• Advanced dashboard has a dot in the uuid #1112
• Average read latency is miscalculated #1110
• Manager showing "offline" despite metrics being present #1109
• branch-3.5 grafana failure #1104
• Alternator: "Average UpdateItem latency by Instance", show data in seconds and not milliseconds #1101
• Alternator dashboard doesn't have GetItem Latencies #1100

Known issues

Following Scylla-Monitoring 2.2 ports change, Prometheus will listen to both old port and the new to help during the migration.
This is was found to cause issues when the port in scylla_manager_server.yml is changed.

We suggest that following a Scylla-Manager upgrade to version 2.2, edit prometheus/prometheus.yml.template and remove scylla_manager1 job from the file.

Branch 3.5

New in Branch 3.5

• Add the dashboard id when submitting an issue enhancement #1031
• Upgrade to grafana 7.1.5 #1030
• Support Manager 2.2 ports enhancement #1010
• Add multi-version support to docs documentation enhancement #1003
• Point from the docs to older version in doc sites documentation enhancement #1001
• Update the dashboard header parts enhancement #970
• remove Scylla os 3.1 and 3.2, manager 1.3, 1.4 #960
• Clearer separation between User and Advance dashboard #947
• Please add alert for backup failure. #933
• Add manager information to the overview dashboard
• need dashboard for Manager 2.2 #924
• Add support for LDAP auth in grafana #912
• use rows with alternator #909
• Adding a grafana-image-renderer to scylla-monitoring #904
• Add Scylla datasource to grafana

Bug fixes

• Dashboard multiple, Stat Joining bug #976
• switch all irate to rate #958
• Scylla Manager progress panel should be percent bug #968
• Prometheus version downgrade due to Consul API issue

Branch 3.4.3

• Add Podman support