So, Range latency(or Scan in case of coordinator) appears only when we are reading multiple partition key in the same query?
пт, 7 июл. 2017 г. в 20:29, Chris Lohfink <clohfin...@gmail.com>: > The coordinator read/scan (Scan is just different naming for the Range, so > coordinator view of RangeLatency) is the latencies from the coordinator > perspective, so it includes network latency between replicas and such. This > which is actually added for speculative retry (why there is no > coordinatorWriteLatency). Only the CoordinatorReadLatency is used for it > however. > > The Read/RangeLatency metrics are for local reads, basically just how long > to read from disk and merge with sstables. > > The View* metrics are only relevant to materialized views. There actually > is a partition lock for updates which ViewLockAcquireTime gives > visibility too. Also there are sometimes reads required for updating > materialized views, which ViewReadTime is for tracking. For more details id > recommend > https://opencredo.com/everything-need-know-cassandra-materialized-views/ > > Chris > > On Fri, Jul 7, 2017 at 9:42 AM, ZAIDI, ASAD A <az1...@att.com> wrote: > >> What exactly does mean CoordinatorScanLatency for example >> >> CoordinatorScanLatency is a timer metric that present coordinator range >> scan latency for table. >> >> Is it latency on full table scan or maybe range scan by clustering key? >> >> It is range scan.. clustering key is used to only store >> data in sorted fashion – partition key along with chosen partitioner helps >> in range scan of data. >> >> Can anybody write into partition while locked? >> >> Writes are atomic – it depends on your chosen consistency >> level to determine if writes will fail or succeed. >> >> >> >> *From:* Павел Сапежко [mailto:amelius0...@gmail.com] >> *Sent:* Friday, July 07, 2017 8:23 AM >> *To:* user@cassandra.apache.org >> *Subject:* Re: Understanding of cassandra metrics >> >> >> >> Are you really think that I don't read docs? Do you have enough >> information in the documentation? I think no. What exactly does mean >> CoordinatorScanLatency >> for example? Is it latency on full table scan or maybe range scan by >> clustering key? What exactly mean ViewLockAcquireTime? What is >> "partition lock"? Can anybody write into partition while locked? Etc. >> >> пт, 7 июл. 2017 г. в 13:01, Ivan Iliev <ivan.iliev.il...@gmail.com>: >> >> 1st result on google returns: >> >> >> >> http://cassandra.apache.org/doc/latest/operating/metrics.html >> <https://urldefense.proofpoint.com/v2/url?u=http-3A__cassandra.apache.org_doc_latest_operating_metrics.html&d=DwMFaQ&c=LFYZ-o9_HUMeMTSQicvjIg&r=FsmDztdsVuIKml8IDhdHdg&m=xLXSLhev337PLD18_DQUB610StqO2URoRumDH-UXdtk&s=xTDUr1_OtTvWNi15F4GNSYs6lzQHi-2wX7oa2T37RCY&e=> >> >> >> >> On Fri, Jul 7, 2017 at 12:16 PM, Павел Сапежко <amelius0...@gmail.com> >> wrote: >> >> Hello, I have several question about cassandra metrics. What does exactly >> mean the next metrics: >> >> - CoordinatorReadLatency >> - CoordinatorScanLatency >> - ReadLatency >> - RangeLatency >> - ViewLockAcquireTime >> - ViewReadTime >> >> -- >> >> С уважением, >> >> Павел Сапежко >> >> skype: p.sapezhko >> >> >> >> -- >> >> С уважением, >> >> Павел Сапежко >> >> skype: p.sapezhko >> > > -- С уважением, Павел Сапежко skype: p.sapezhko
