storage-schemas.conf

# This config file sets up your retention rules.
# It is an extension of http://graphite.readthedocs.io/en/latest/config-carbon.html#storage-schemas-conf
# Note:
# * You can have 0 to N sections
# * The first match wins, starting from the top. If no match found, we default to single archive of minutely points, retained for 7 days in 2h chunks
# * The patterns are unanchored regular expressions, add '^' or '$' to match the beginning or end of a pattern.
# * When running a cluster of metrictank instances, all instances should have the same agg-settings.
# * Unlike whisper (graphite), the config doesn't stick: if you restart metrictank with updated settings, then those
# will be applied. The configured rollups will be saved by primary nodes and served in responses if they are ready.
# (note in particular that if you remove archives here, we will no longer read from them)
# * Retentions must be specified in order of increasing interval and retention
# * The reorderBuffer an optional buffer that temporarily keeps data points in memory as raw data and allows insertion at random order. The specified value is how many datapoints, based on the raw interval specified in the first defined retention, should be kept before they are flushed out. This is useful if the metric producers cannot guarantee that the data will arrive in order, but it is relatively memory intensive. If you are unsure whether you need this, better leave it disabled to not waste memory. When enabled, you can optionally via 'reorderBufferAllowUpdate' allow updating the value of data points already received (if the timestamp falls within the reorder buffer window).
#
# A given rule is made up of at least 3 lines: the name, regex pattern, retentions and optionally the reorder buffer size.
# The retentions line can specify multiple retention definitions. You need one or more, space separated.
#
# There are 2 formats for a single retention definition:
# 1) 'series-interval:count-of-datapoints'                   legacy and not easy to read
# 2) 'series-interval:retention[:chunkspan:numchunks:ready]' more friendly format with optionally 3 extra fields
#
#Series intervals and retentions are specified using the following suffixes:
#
#s - second
#m - minute
#h - hour
#d - day
#y - year
#
# The final 3 fields are specific to metrictank and if unspecified, use sane defaults.
# See https://github.com/grafana/metrictank/blob/master/docs/memory-server.md for more details
#
# chunkspan: duration of chunks. e.g. 10min, 30min, 1h, 90min...
# must be valid value as described here https://github.com/grafana/metrictank/blob/master/docs/memory-server.md#valid-chunk-spans
# Defaults to a the smallest chunkspan that can hold at least 100 points.
#
# numchunks: number of raw chunks to keep in in-memory ring buffer
# See https://github.com/grafana/metrictank/blob/master/docs/memory-server.md for details and trade-offs, especially when compared to chunk-cache
# which may be a more effective method to cache data and alleviate workload for cassandra.
# Defaults to 2
#
# ready: whether, or as of what data timestamp, the archive is ready for querying.
# This is useful if you recently introduced a new archive, but it's still being populated, so you want to control whether (or to which extent) the archive can be used for queries.
# It supports two syntaxes:
# * unix timestamp: the archive contains data as of this timestamp
# * boolean: (legacy): whether or not the archive is completely ready or not ready at all.
# Defaults to true
#
# Here's an example with multiple retentions:
# [apache_busyWorkers]
# pattern = ^servers\.www.*\.workers\.busyWorkers$
# retentions = 1s:1d:10min:1,1m:21d,15m:5y:2h:1:false
#
# This example has 3 retention definitions, the first and last override some default options (to use 10minutely and 2hourly chunks and only keep one of them in memory
# and the last rollup is marked as not ready yet for querying.

[default]
pattern = .*
retentions = 1s:35d:10min:7
# reorderBuffer = 20
# reorderBufferAllowUpdate = true