pinba-influxer

InfluxDB Preparation

Create database:

CREATE DATABASE "pinba"

Create retention policy:

CREATE RETENTION POLICY "realtime" ON "pinba" DURATION 2h REPLICATION 1 DEFAULT
CREATE RETENTION POLICY "month" ON "pinba" DURATION 4w REPLICATION 1
CREATE RETENTION POLICY "year" ON "pinba" DURATION 52w REPLICATION 1
CREATE RETENTION POLICY "infinity" ON "pinba" DURATION INF REPLICATION 1

And now, interesting part - continuous queries.

For downsampling realtime data to 10s:

CREATE CONTINUOUS QUERY "cq_requests_10s" ON "pinba" BEGIN
    SELECT
        count("request_time") / 10 AS "rps",
        percentile("request_time", 25) AS "p25",
        percentile("request_time", 75) AS "p75",
        percentile("request_time", 95) AS "p95",
        max("request_time") AS "max"
    INTO "month"."requests"
    FROM "requests"
    GROUP BY time(10s), * END

CREATE CONTINUOUS QUERY "cq_timers_10s" ON "pinba" BEGIN
    SELECT
        sum("hits") / 10 AS "rps",
        percentile("value", 25) AS "p25",
        percentile("value", 75) AS "p75",
        percentile("value", 95) AS "p95",
        max("value") AS "max"
    INTO "month"."timers"
    FROM "timers"
    GROUP BY time(10s), * END

For downsampling realtime data to 60s:

CREATE CONTINUOUS QUERY "cq_requests_60s" ON "pinba" BEGIN
    SELECT
        count("request_time") / 60 AS "rps",
        percentile("request_time", 25) AS "p25",
        percentile("request_time", 75) AS "p75",
        percentile("request_time", 95) AS "p95",
        max("request_time") AS "max"
    INTO "year"."requests"
    FROM "requests"
    GROUP BY time(60s), * END

CREATE CONTINUOUS QUERY "cq_timers_60s" ON "pinba" BEGIN
    SELECT
        sum("hits") / 60 AS "rps",
        percentile("value", 25) AS "p25",
        percentile("value", 75) AS "p75",
        percentile("value", 95) AS "p95",
        max("value") AS "max"
    INTO "year"."timers"
    FROM "timers"
    GROUP BY time(60s), * END

And for downsampling realtime data to 10m:

CREATE CONTINUOUS QUERY "cq_requests_10m" ON "pinba" BEGIN
    SELECT
        count("request_time") / 600 AS "rps",
        percentile("request_time", 25) AS "p25",
        percentile("request_time", 75) AS "p75",
        percentile("request_time", 95) AS "p95",
        max("request_time") AS "max"
    INTO "year"."requests"
    FROM "requests"
    GROUP BY time(10m), * END

CREATE CONTINUOUS QUERY "cq_timers_10m" ON "pinba" BEGIN
    SELECT
        sum("hits") / 600 AS "rps",
        percentile("value", 25) AS "p25",
        percentile("value", 75) AS "p75",
        percentile("value", 95) AS "p95",
        max("value") AS "max"
    INTO "year"."timers"
    FROM "timers"
    GROUP BY time(10m), * END