The article is very well written. Their team do leverage great tooling for thing such as Gor. I love Gor. I used it + Mongo OpsLog when migrate from AWS -> DigitalOcean and eventually have to migrate back to AWS because DigitalOcean cannot deliver good performance as AWS and they throtle CPU(lots of steal CPU).
OpsManager is great for team that don't have dedicated devops I think and have great dashboard/visualization.
However, run your own MongoDB is very easy. Not like Postgres(unless you used RDS). However, when using RDS, you still have downtime when upgrading db, it still have a small amount of time the standby in MultiAZ is promoted to master, DNS is updated, and during that yourcurrent primary is not writeable or even not available. MongoDB is way easier to operator. You add/delete/node on fly and client auto discover network topology. Plus https://docs.mongodb.com/manual/administration/production-no... this links give great info to tune it: thing like run on XFS, mount with notatime options...puts opslog on high iops volume etc...
Peformance isn't a factor to pick your database much nowsaday. They looks great on the benchmark. However, try for yourself before pick one on your workload. Postgres does has its own ward.
Pick a database based on how well your team confident with it, how does the database help you move fast enough or deliver business value. Don't follow the hype or silly benchmark.
OpsManager is great for team that don't have dedicated devops I think and have great dashboard/visualization.
However, run your own MongoDB is very easy. Not like Postgres(unless you used RDS). However, when using RDS, you still have downtime when upgrading db, it still have a small amount of time the standby in MultiAZ is promoted to master, DNS is updated, and during that yourcurrent primary is not writeable or even not available. MongoDB is way easier to operator. You add/delete/node on fly and client auto discover network topology. Plus https://docs.mongodb.com/manual/administration/production-no... this links give great info to tune it: thing like run on XFS, mount with notatime options...puts opslog on high iops volume etc...
Peformance isn't a factor to pick your database much nowsaday. They looks great on the benchmark. However, try for yourself before pick one on your workload. Postgres does has its own ward.
Pick a database based on how well your team confident with it, how does the database help you move fast enough or deliver business value. Don't follow the hype or silly benchmark.