[SERVER-5026] Fast, thread-safe counters for statistics. Created: 21/Feb/12 Updated: 25/Aug/23 Resolved: 19/Apr/22 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Concurrency, Performance |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | New Feature | Priority: | Minor - P4 |
| Reporter: | Andy Schwerin | Assignee: | Vojislav Stojkovic |
| Resolution: | Declined | Votes: | 1 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||||||
| Sprint: | Service Arch 2022-05-02 | ||||||||
| Participants: | |||||||||
| Description |
|
For performance measurement, we should have a system of counters that are not lossy, and whose writes do not impact performance (i.e., increase cache misses or lock contention). |
| Comments |
| Comment by Lauren Lewis (Inactive) [ 21/Dec/21 ] |
|
We haven’t heard back from you in at least 1 year, so I'm going to close this ticket. If this is still an issue for you, please provide additional information and we will reopen the ticket. |
| Comment by Kyle Suarez [ 24/Mar/16 ] |
|
I'm assuming we're talking about operation counters like the OpCounters? If so, it would be nice if the counters were 64-bit integers, which would then kill Another related issue that affects the current OpCounters is that all of the counters fit on the same cache line, which could lead to contention. (This is just a "what-if" and should be something we keep in mind when doing performance tests.) |