New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Issue #167 #207
Issue #167 #207
Conversation
Fixed up file and directory handling in response to issues graphite-project#17 and graphite-project#18 Conflicts: lib/carbon/writer.py
…xisting to prevent log spam when new metrics are created in existing directory trees Cleaned up whitespace in log.py
Stop log spam, clean up some more formatting.
Token bucket based rate limiting deque instead of list for datapoint storage inside cache removed locks and need for locks during threaded interaction between writer and cache MetricCache.pop now automatically returns the metric with the largest number of cached datapoints No more cache sorts!
Allow setting a custom umask
Configurable cache writer strategies defaultdict for the main cache datastructure (backwards compatible with 2.4 even!) some other stuff I've probably forgotten!
…ter explain the file creation options
small whitespace fix
…uency carbon.conf: Add [aggregator] WRITE_BACK_FREQUENCY
Refactor cache to not require locks Add different write ordering behaviors Use token buckets for write rate limiting Add FORWARD_ALL parameter to aggregator Refactor hashing for clarity
improve CH routing performance when replfactor=1 Re-merge due to loss in a big merge commit
The MAX value is an integer, but was being divided by 60, and so was producing an integer result. Therefore, if MAX was < 60, the result is 0 and no new metrics are created after the initial MAX is hit. If MAX is >= 60, new metrics will be created, though not at exactly the rate that would be expected. The TokenBucket expects floating point values, so force the division to be floating point. Fixes issue graphite-project#55.
Treat MAX_CREATES_PER_MINUTE as a float.
Fix misreporting cache size/queues metrics
Add min & max aggregation rules
otherwise they goes as "nan" in the whisper files
amqp: ignore NaN values
…or datapoints of a set of metrics, which gives much better performance than single request per metric.
Add new type of cache query - 'cache-query-bulk'
Updated example config to demonstrate new options
Added Configurable files for aggregation and rewrite rules
Fixes compat with twisted 13.2.0 by removing dead code.
Fixed broken aggregation rules file config
Can you close this out, and re-apply your patch on top of the 0.9.x branch without all of the intermediate commits that you've added to this PR? Thanks! |
i've not added intermediate commits, github is showing you the diff against On Mon, Jan 27, 2014 at 5:19 PM, Peter N notifications@github.com wrote:
|
Sorry, this PR is a total mess. You rebased off |
This is a patch/pull request for 0.9.x, not master. resolves issue 167