3 ms·

Hi! The performance gap between questDB on both ingestion and queries versus timescale or influx is very significant (orders of magnitude faster). questDB is be

by j1897 6y ago

Hi! The performance gap between questDB on both ingestion and queries versus timescale or influx is very significant (orders of magnitude faster). questDB is being used in production for critical use cases at large companies and is definitively stable. Watch out for our new release next week

g8oz 6y ago
Any performance metric comparisons available?
- j1897 6y ago
  We ran this experiment Vs influx some time ago for ingestion here https://medium.com/questdb/speeding-up-influx-line-protocol-c13e764f444d https://medium.com/questdb/speeding-up-influx-line-protocol-... In terms of queries, see our previous post on HN to get a sense https://news.ycombinator.com/item?id=22803504 https://news.ycombinator.com/item?id=22803504
  - sgt 6y ago
    Ingestion rate is pretty good in Influx and we are happy with that for now. If QuestDB is better I definitely won't complain. The issues with Influx is data integrity and also the difficulty in manipulating time series data if that becomes an issue (e.g turning all "0" string values into a 0i). How does QuestDB fare in this regard?
    - bluestreak 6y ago
      We take data integrity very seriously. This is one of the reasons QuestDB actually exists. We haven't had reports of data corruption or loss yet. We can commit durably if needed too. To manipulate data we support "add" and "delete" column on the fly. We can also add column replace and type change if needed. This is pretty easy to do. PostgreSQL wire is in beta. It works with JDBC driver and we will add metadata support quite soon.
sgt 6y ago
I guess I was more interested in finding out whether questDB is mature for production use. The product is impressive, but it looks like the documentation has some way to go. This is often indicative of the maturity of the product itself. I was very interested in the Postgres wire protocol and its usage. This part does not seem to be very mature. Examples: * The documentation doesn't mention a word of the default username/password on port 8812. I had to find this on a mailing list somewhere after 10 minutes of googling. Has this been hardcoded into the code? * If you were to connect a typical Postgres tool this port, you'd be able to do selects but not even see the tables. If using psql and doing \d I get: ERROR: Invalid column: oid

by j1897 6y ago

g8oz 6y ago

Any performance metric comparisons available?

j1897 6y ago
We ran this experiment Vs influx some time ago for ingestion here https://medium.com/questdb/speeding-up-influx-line-protocol-c13e764f444d https://medium.com/questdb/speeding-up-influx-line-protocol-... In terms of queries, see our previous post on HN to get a sense https://news.ycombinator.com/item?id=22803504 https://news.ycombinator.com/item?id=22803504
- sgt 6y ago
  Ingestion rate is pretty good in Influx and we are happy with that for now. If QuestDB is better I definitely won't complain. The issues with Influx is data integrity and also the difficulty in manipulating time series data if that becomes an issue (e.g turning all "0" string values into a 0i). How does QuestDB fare in this regard?
  - bluestreak 6y ago
    We take data integrity very seriously. This is one of the reasons QuestDB actually exists. We haven't had reports of data corruption or loss yet. We can commit durably if needed too. To manipulate data we support "add" and "delete" column on the fly. We can also add column replace and type change if needed. This is pretty easy to do. PostgreSQL wire is in beta. It works with JDBC driver and we will add metadata support quite soon.

sgt 6y ago

I guess I was more interested in finding out whether questDB is mature for production use. The product is impressive, but it looks like the documentation has some way to go. This is often indicative of the maturity of the product itself. I was very interested in the Postgres wire protocol and its usage. This part does not seem to be very mature. Examples: * The documentation doesn't mention a word of the default username/password on port 8812. I had to find this on a mailing list somewhere after 10 minutes of googling. Has this been hardcoded into the code? * If you were to connect a typical Postgres tool this port, you'd be able to do selects but not even see the tables. If using psql and doing \d I get: ERROR: Invalid column: oid