> Data durability > > As a data service our first priority is keeping your data safe. We utilize WAL-E (link is external), the popular continuous archiving to

> Data durability > > As a data service our first priority is keeping your data safe. We utilize WAL-E (link is external), the popular continuous archiving tool for Postgres. A slight aside, but it's definitely worth looking at Postgres' WAL system [1], which is very cool technology that recently got better in 9.4 in that it can stream a "logical" representation of the WAL instead of the previous format which was largely only good for internal use. The WAL-E project that persists WAL to S3 was originally started for Heroku Postgres, but is now in widespread use elsewhere including the new Citus initiative. Anyway, I'm a little biased because this was bootstrapped by ex-colleagues, but I'm excited to see an alternative to Redshift (not necessarily the target competitor here, but one interesting technology that Citus could viably replace) that allows for more consistent query performance and has all the modern features of Postgres (i.e. as opposed to being permanently locked into 8.0.2 with Redshift). [1] http://www.postgresql.org/docs/current/static/wal-intro.html http://www.postgresql.org/docs/current/static/wal-intro.html [2] http://www.postgresql.org/docs/9.4/static/logicaldecoding-explanation.html http://www.postgresql.org/docs/9.4/static/logicaldecoding-ex... [3] https://github.com/wal-e/wal-e https://github.com/wal-e/wal-e