Surface the transaction retry chart more prominently in the DB Console and add documentation and tool tips on the Distributed metric dashboard #65856
Comments
kevin-v-ngo
added
C-enhancement
kevin-v-ngo
added
the
A-sql-observability
|
We already export a metric txn.restarts as well as a bunch of reason-specifc metrics like txn.restarts.writetooold. These are shown in the KV Transaction Restarts chart on the distributed dashboard in the web console. Are these not what you want? |
|
Ah thanks for pointing this out Ben! Did not know this was here 😀 It would be good if this was more discoverable. Are there any differences between "KV" transactions and how we refer to SQL transactions on the SQL dashboard page? |
I may be misremembering but I think we used to have a transactions section of the dashboard (which apparently got folded into "distributed"?).
There are some, so you won't get the numbers to line up one-to-one. For example, there are some internal KV transactions that weren't initiated by SQL, and when a transaction auto-retries I think one SQL transaction may get counted as two KV transactions (I think it varies depending on exactly why there was a retry). But it's close enough that you can generally ignore the differences. For questions like "am I experiencing retries? and if so, why?" these metrics are what you want. |
exalate-issue-sync
bot
changed the title
exalate-issue-sync
bot
changed the title
|
Adding @Annebirzin and @ianjevans for visibility. @Annebirzin, the SQL activity page might be a good place for this when we introduce a metrics tab. |
|
Pinging @taroface. I'm having trouble finding any docs release note issues for the new metrics dashboards. I think we're missing docs for the Distributed, Queues, and Slow Queries dashboards, but I think that's because there never was a release note string that triggered a doc issue. I don't know why that occurred, though. |
|
So after some digging by Jordan, it appears that the missing dashboards have been in the codebase for quite some time, but have been undocumented. They likely were added before our RN text -> doc issue process began, so that's why we don't have doc issues for them. I'll create a separate issue to handle adding the docs for these dashboards. |
|
To update this thread, we are moving the transaction restart chart (broken down by errors) into the SQL dashboard beneath transactions to make this more discoverable. @ianjevans, FYI, we should provide documentation on what to do for each error. I think this many are already captured in this reference documentation: https://www.cockroachlabs.com/docs/stable/transaction-retry-error-reference.html |
66973: ui: surface the transaction restarts chart r=matthewtodd a=matthewtodd Resolves #65856 Release note (ui change): The KV transaction restarts chart was moved from the "distributed" metrics to the "sql" metrics page so as to be close to the SQL transactions chart, for more prominent visibility. 66979: opt: add cost penalty for scans with large cardinality r=rytaft a=rytaft **opt: ensure we prefer a reverse scan to sorting a forward scan** This commit fixes an issue where in some edge cases the optimizer would prefer sorting the output of a forward scan over performing a reverse scan (when there is no need to sort the output of the reverse scan). Release note (performance improvement): The optimizer now prefers performing a reverse scan over a forward scan + sort if the reverse scan eliminates the need for a sort and the plans are otherwise equivalent. This was the case before in most cases, but some edge cases with a small number of rows have been fixed. **opt: add cost penalty for scans with large cardinality** This commit adds a new cost function, `largeCardinalityRowCountPenalty`, which calculates a penalty that should be added to the row count of scans. It is non-zero for expressions with unbounded maximum cardinality or with maximum cardinality exceeding the row count estimate. Adding a few rows worth of cost helps prevent surprising plans for very small tables or for when stats are stale. Fixes #64570 Release note (performance improvement): When choosing between index scans that are estimated to have the same number of rows, the optimizer now prefers indexes for which it has higher certainty about the maximum number of rows over indexes for which there is more uncertainty in the estimated row count. This helps to avoid choosing suboptimal plans for small tables or if the statistics are stale. Co-authored-by: Matthew Todd <todd@cockroachlabs.com> Co-authored-by: Rebecca Taft <becca@cockroachlabs.com>
We have a Transaction Restarts chart in the DB console under the distributed dashboard. Currently there is no documentation on this part of the console and we should also provide a way to surface this transaction restart view more prominently. One place to consider is the SQL dashboard where there are related counters for statements and transactions.
Documentation we should also update: https://www.cockroachlabs.com/docs/v21.1/ui-overview.html
Epic CRDB-6570
The text was updated successfully, but these errors were encountered: