TimescaleDB VS promscale

Don't forget PostgreSQL extensions. For something like a chat log, TimescaleDB (https://www.timescale.com/) can be surprisingly efficient. It will handle partitioning for you, with additional features like data reordering, compression, and retention policies.

Offboard it to Postgres specialists like https://www.timescale.com/

TimescaleDB{:target="_blank"}: an extension of PostgreSQL that adds time-series capabilities to the relational database model. It provides scalability and performance optimizations for handling large volumes of time-stamped data while maintaining the flexibility of a relational database.

Have some experience with https://www.timescale.com in this context

What about getting in touch with commercial organisations that have products/services based on PostgreSQL? For example Timescale, EDB, and Citus Data, or really any hosting provider that offers a managed PostgreSQL service.

There is Timescale.

Friends don't let their friends choose Mysql :)

A super long time ago (decades) when I was using Oracle regularly I had to make a decision on which way to go. Although Mysql then had the mindshare I thought that Postgres was more similar to Oracle, more standards compliant, and more of a real enterprise type of DB. The rumor was also that Postgres was heavier than MySQL. Too many horror stories of lost data (MyIsam), bad transactions (MyIsam lacks transaction integrity), and the number of Mysql gotchas being a really long list influenced me.

In time I actually found out that I had underestimated one of the most important attributes of Postgres that was a huge strength over Mysql: the power of community. Because Postgres has a really superb community that can be found on Libera Chat and elsewhere, and they are very willing to help out, I think Postgres has a huge advantage over Mysql. RhodiumToad [Andrew Gierth] https://github.com/RhodiumToad & davidfetter [David Fetter] https://www.linkedin.com/in/davidfetter are incredibly helpful folks.

I don't know that Postgres' licensing made a huge difference or not but my perception is that there are a ton of 3rd party products based on Postgres but customized to specific DB needs because of the more liberalness of the PG license which is MIT/BSD derived https://www.postgresql.org/about/licence/

Some of the PG based 3rd party DBs:

Enterprise DB https://www.enterprisedb.com/ - general purpose PG with some variants

Greenplum https://greenplum.org/ - Data warehousing

Crunchydata https://www.crunchydata.com/products/hardened-postgres - high security Postgres for regulated environments

Citus https://www.citusdata.com - Distributed DB & Columnar

Timescale https://www.timescale.com/

Why Choose PG today?

If you want better ACID: Postgres

If you want more compliant SQL: Postgres

If you want more customizability to a variety of use-cases: Postgres using a variant

If you want the flexibility of using NOSQL at times: Postgres

If you want more product knowledge reusability for other backend products: Postgres

TimescaleDB is Postgres with extensions to automatically partition tables for fast processing of time series data.

Now that Promscale has been deprecated, what are the other ideal means of self-hosted long term Prometheus storage?

Oh wow, I browsed the project just a few weeks ago, didn't see it then. I see the deprecation is recent (https://github.com/timescale/promscale/issues/1836)

They say:

> if you want to have a seamless experience between metrics and traces, then current experience of stitching together Prometheus & Jaeger is not great.

But I wonder if using Promscale https://github.com/timescale/promscale would make Prometheus & Jaeger not such a big problem as SigNoz imply.

Promscale readme:

> Promscale is a unified metric and trace observability backend for Prometheus, Jaeger and OpenTelemetry built on PostgreSQL and TimescaleDB.

Either way, SigNoz seems interesting indeed. And am glad to see that SigNoz supports OpenTelemetry.

Hi! So the team is over 100 at this point, but engineering effort is spread across multiple products at this point.

The core timescaledb repo [0] has 10-15 primary engineers (although we are aggressively hiring for database internal engineers), with a few others working on DB hyperfunctions and our function pipelining [1] in a separate extension [2]. I think generally the set of folks who contribute to low-level database internals in C is just smaller than other type of projects.

We also have our promscale product [3], which is our observability backend powered by SQL & TimescaleDB.

And then there is Timescale Cloud, which is obviously a large engineering effort (most of which does not happen in public repos).

And we are hiring. Fully remote & global.

https://www.timescale.com/careers

[0] https://github.com/timescale/timescaledb

[1] https://www.timescale.com/blog/function-pipelines-building-f...

[2] https://github.com/timescale/timescaledb-toolkit

[3] https://github.com/timescale/promscale ; https://github.com/timescale/tobs

Promscale is a connector for Prometheus, one of the leading open-source monitoring solutions. Promscale is developed by Timescale, a time series database with full compatibility to Postgres. Since logs are time series events, Timescale developed Promscale to ingest events from Prometheus and make them available in SQL. You can install Promscale in numerous ways.

In case it helps, Promscale (from Timescale) offers long-term storage for Prometheus data and supports both PromQL and SQL queries. Here's the project page: https://www.timescale.com/promscale/ and the repo is here https://github.com/timescale/promscale It also support OpenTelemetry tracing if that's of interest.

At first, let's give the definition of `time series`. This is a series of (timestamp, value) pairs ordered by timestamp. The `value` may contain arbitrary data - a floating-point value, a text, a json, a data structure with many columns, etc. Each time series is uniquely identified by its name plus an optional set of {label="value"} labels. For example, temperature{city="London",country="UK"} or log_stream{host="foobar",datacenter="abc",app="nginx"}.

ClickHouse is perfectly optimized for storing and querying of such time series, including metrics. That's true that ClickHouse isn't optimized for handling millions of tiny inserts per second. It prefers infrequent batches with big number of rows per each batch. But this isn't the real problem in practice, because:

1) ClickHouse provides Buffer table engine for frequent inserts.

2) It is easy to create a special proxy app or library for data buffering before sending it to ClickHouse.

TimescaleDB provides Promscale [1] - a service, which allows using TimescaleDB as a storage backend for Prometheus. Unfortunately, it doesn't show outstanding performance comparing to Prometheus itself and to other remote storage solutions for Prometheus. Promscale requires more disk space, disk IO, CPU and RAM according to production tests [2], [3].

[1] https://github.com/timescale/promscale

[2] https://abiosgaming.com/press/high-cardinality-aggregations/

[3] https://valyala.medium.com/promscale-vs-victoriametrics-reso...

Full disclosure: I'm CTO at VictoriaMetrics - competing solution for TimescaleDB. VictoriaMetrics is built on top of architecture ideas from ClickHouse.

Promscale + TimescaleDB