Hi all, I've taken up my work on exemplars again; the above PR and design doc are still relevant but I'll paste the links again below.
I've changed the PR from a draft to open, it's still rough but I would appreciate feedback. Note that the implementation is in memory, using a single fixed size circular buffer (size configurable by users) so that overall memory usage by exemplar storage is bounded. design doc <https://docs.google.com/document/d/1ymZlc9yuTj8GvZyKz1r3KDRrhaOjZ1W1qZVW_5Gj7gA/edit?usp=sharing> PR <https://github.com/prometheus/prometheus/pull/6635> Thanks, Callum. On Fri, Mar 20, 2020 at 3:31 AM Bartłomiej Płotka <[email protected]> wrote: > Awesome! Exciting work, thanks for sharing. Will take a look in following > days / weekend (: > > Kind Regards, > Bartek > > On Thu, 19 Mar 2020 at 21:27, Callum Styan <[email protected]> wrote: > >> Hi all, >> >> For those who are interested, I've added notes on how the exemplar >> storage might handle tail based sampling of traces. These notes are based >> on discussions I've had with a few people involved in/working on tracing >> systems. >> >> As usual, please let me know if you have any feedback. >> >> The current exemplar storage PR is here: >> https://github.com/prometheus/prometheus/pull/6635 In the current state, >> building Prometheus from that branch will allow you to scrape, store, and >> query exemplars. >> >> Thanks, >> Callum. >> >> On Thu, Jan 9, 2020 at 3:39 PM Callum Styan <[email protected]> >> wrote: >> >>> Hi all, >>> >>> As many of you know we've been planning to add minimal in-memory storage >>> of exemplars within Prometheus for some time now. Some work has already >>> been done in this area; including the exposition of exemplars in the Python >>> client >>> <https://github.com/prometheus/client_python/search?q=exemplar&unscoped_q=exemplar>, >>> and parsing of the Open Metrics exemplar format >>> <https://github.com/prometheus/prometheus/pull/6292> by the >>> M3/Chronosphere folks, who also proposed a possible storage interface >>> here <https://github.com/prometheus/prometheus/pull/6309>. >>> >>> More recently some of us have been putting together a design doc for the >>> end to end exemplar experience within Prometheus; discussing what the query >>> flow for exemplars might look like from Grafana itself, what the storage >>> interface (external and internal) API might look like, and how the internal >>> storage could be implemented. >>> >>> Some of you have also seen the experimental implementation PR >>> <https://github.com/prometheus/prometheus/pull/6574> I opened a few >>> days ago. From the comments there it sounds like there may be exemplar >>> querying use cases that haven't been discussed yet, which I would ask you >>> to make note of in the design doc. >>> >>> With all that said, here's the design doc >>> <https://docs.google.com/document/d/1ymZlc9yuTj8GvZyKz1r3KDRrhaOjZ1W1qZVW_5Gj7gA/edit?usp=sharing>. >>> Note that an implementation has not been finalized yet. >>> >>> Thanks, >>> Callum. >>> >> -- >> You received this message because you are subscribed to the Google Groups >> "Prometheus Developers" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/prometheus-developers/CAN2d5OT%2BJi4DB4fupTyjsgRa_T%2ByZiiVfG-FuEJDRV%3DWXDJDvg%40mail.gmail.com >> <https://groups.google.com/d/msgid/prometheus-developers/CAN2d5OT%2BJi4DB4fupTyjsgRa_T%2ByZiiVfG-FuEJDRV%3DWXDJDvg%40mail.gmail.com?utm_medium=email&utm_source=footer> >> . >> > -- You received this message because you are subscribed to the Google Groups "Prometheus Developers" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-developers/CAN2d5OS_9wamDTbEdJReYdNtuY81oZTkHAS4gCvBf%3Dm0jbWB9A%40mail.gmail.com.
