guillotina-rediscache 2.1.3

guillotina cache implementation using redis + lru in-memory cache

guillotina_rediscache implements redis into guillotina with an additional in-memory layer cache.

In order to coordinate invalidating the in-memory cache, guillotina_rediscache utilizes the pub/sub feature redis provides.

Configuration

app_settings for this:

{
  "databases": {
    "db": {
      ...
      "cache_strategy": "redis"
      ...
    }
  },
  "redis": {
      'host': 'localhost',
      'port': 6379,
      'ttl': 3600,
      'memory_cache_size': 209715200,
      'pool': {
          'minsize': 5,
          'maxsize': 100
      }
  }
}

Run measures

Using guillotina run command:

./bin/g run --script=measures/serialize.py -c measures/config.yaml

With profiling:

./bin/g run --script=measures/serialize.py -c measures/config.yaml --line-profiler --line-profiler-matcher="*serialize*"

2.1.3 (2019-06-18)

• Restrict version to guillotina < 5 [vangheem]

2.1.2 (2019-06-12)

• Use expire param with setting redis cache value on file data manager [vangheem]

2.1.1 (2019-01-18)

• When applying invalidations, make sure to not allow another starting transaction to conflict with asyncio invalidation task [vangheem]
• Fix endpoint @redis-clear-cache [masipcat]

2.1.0 (2018-11-20)

• Upgrade to Guillotina >= 4.3 [bloodbare]

2.0.4 (2018-10-06)

• Do not fail memory cache if not configured [vangheem]

2.0.3 (2018-07-20)

• Remove old guillotina compatibility for py 3.7 [bloodbare]

2.0.2 (2018-06-18)

• Improve size calculation [jordic]

2.0.1 (2018-06-15)

• Change default cache size to 209715200 [vangheem]

2.0.0 (2018-06-15)

• Cache based on memory size instead of total number of keys [jordic]

1.3.4 (2018-04-03)

• Fix fixture [vangheem]

1.3.3 (2018-04-03)

• Upgrade testing for guillotina 3.0 [vangheem]

1.3.2 (2018-03-26)

• Use after commit hook to update data for file manager implementation [vangheem]

1.3.1 (2018-03-19)

• Fix redis file manager finish method [vangheem]

1.3.0 (2018-03-19)

• Provide redis file manager [vangheem]

1.2.0 (2018-03-14)

• Upgrade to work with guillotina 2.4.x [vangheem]

1.1.6 (2018-03-01)

• Handle errors while canceling task init task [vangheem]

1.1.5 (2018-03-01)

• Handle errors while canceling task [vangheem]

1.1.4 (2018-02-15)

• Fix cache stats endpoint [vangheem]

1.1.3 (2018-01-22)

• Be able to disable deleting group of cache keys together with cluster_mode option. [vangheem]

1.1.2 (2018-01-17)

• Fix using redis.delete when keys are of length 0 [vangheem]

1.1.1 (2018-01-17)

• batch all cache deletes into one request [vangheem]

1.1.0 (2018-01-12)

• Push cache updates to redis subscriber. This should improve cache hits dramatically [vangheem]

1.0.14 (2018-01-10)

• Only run invalidation task if we have keys to invalidate [vangheem]

1.0.13 (2017-12-15)

• Improve request performance [vangheem]
• Change the way we’re using the redis pool so it reuses connections [vangheem]

1.0.12 (2017-11-30)

• Missing await statement for self.get_redis() [vangheem]

1.0.11 (2017-11-08)

• Handle CancelledError [vangheem]

1.0.10 (2017-11-06)

• upgrade for guillotina 2.0.0 [vangheem]

1.0.9 (2017-10-23)

• Fix handling connection objects and releasing back to pool [vangheem]

1.0.8 (2017-10-23)

• Fix use of pool [vangheem]

1.0.7 (2017-10-23)

• Use pickle instead of json from load/dumps because it is much faster [vangheem]

1.0.6 (2017-10-19)

• Use ujson [vangheem]

1.0.5 (2017-10-02)

• Track all keys needing invalidation and do invalidation in an async task so the request can finish faster. [vangheem]

1.0.4 (2017-05-29)

• Test fixes [vangheem]

1.0.3 (2017-05-26)

• Fix delete not properly invalidating cache [vangheem]

1.0.2 (2017-05-15)

• Fix channel publishing invalidations [vangheem]

1.0.1 (2017-05-15)

• Fix release

1.0.0 (2017-05-15)

• initial release