Until now a custom gzip storage format was used for the collector. This approach is based on the first implementation where a simple JSON dict was exported to a file. In the future, sqlite3 might bring more stability to persistent storage.
Note: Postgres is also a candidate, but I do not want to introduce external dependencies (psychopg2, sqlalchemy) just yet. Better focus on small footprint and stability.
Until now a custom gzip storage format was used for the collector. This approach is based on the first implementation where a simple JSON dict was exported to a file. In the future, sqlite3 might bring more stability to persistent storage.
Note: Postgres is also a candidate, but I do not want to introduce external dependencies (psychopg2, sqlalchemy) just yet. Better focus on small footprint and stability.