Description
During a slow GET investigation for a customer, it was useful to correlate our reporting of slow operations in the memcached log to their application reporting slow requests:
2017-06-28T05:45:34.914126Z WARNING 485: Slow GET operation on connection: 578 ms ([ xxx.xxx.114.151:41590 - xxx.xxx.114.133:11210 ])
|
However, the threshold in the customer environment differs the memcached threshold - they report ops as slow at 100ms, compared to the fixed threshold of 500ms in memcached. We should allow the threshold to be configured to match with different SLAs.
Attachments
Issue Links
- relates to
-
DOC-5546 document how to adjust KV tracing threshold
-
- Closed
-
For Gerrit Dashboard: MB-25632 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
82450,18 | MB-25632: Allow slow op threshold to be customised | master | kv_engine | Status: MERGED | +2 | +1 |
172840,4 | MB-25632: Opcode attributes: document dynamic changing via mcctl | master | kv_engine | Status: MERGED | +2 | +1 |
We probably want this configurable via mcctl or similar (for easy tuning of single memcached process), and also via the memcached config file so ns_server can push down config to the whole cluster.