| HN Mirror

brokentone 2153 days ago

I'm personally working on migrating a DB I've had for tracking some economic time-series data from Influx to TimescaleDB, and I have to say Outflux is the most savage tool name ever.

https://www.outfluxdata.com/

akulkarni 2153 days ago

One of our designers teamed up with one of our web developers, went away for a week, and came back with this website for it:

RedShift1 2153 days ago

It's missing some kind of wavy background thing in the hero section.

mayhemchaos 2153 days ago

Agreed -- every time I tried to query data from the command line I could not get anything to work. I would usually find a piece of working python code and adapt that to run the query instead.

100% opposite of what a CLI interface should be.

Speaking from my own experience, you may save yourself some future effort by moving directly to clickhouse.

Timescale brings its own issues. If your goal is performance, you will be better served by clickhouse.

nwmcsween 2153 days ago

The issue I have with clickhouse is the codebase, it's an absolute behemoth and seemingly embeds musl libc? It also uses a huge amount of SIMD intrinsics for everything when SWAR or really nothing from my view looking in would have been better (memcpy, etc).

caveat: I'm not familiar with the clickhouse codebase. however, usually "embedding" musl libc is about portability - it allows you to build an entirely static binary that can run on practically any box. is this different?

secondly, I don't get where you're coming from faulting a columnar database for using SIMD! why is memcpy better? it comes across as, 'they worked too hard making it fast'!

RedShift1 2153 days ago

Can you elaborate on those issues?

It would be too long. To quickly summarize, from the pain of backups (unless you setup a WAL replica, the load may take your database down), the large size of the data on disk (timescale does offer some compression now, but it's still too much), the low performance of large queries, the memory requirements - it's death by a thousand papercuts!

Don't get me wrong, timescale is a great way to get started with time series - just like sqlite is a great way to get started with databases if all you know is nosql.

However, it quickly brings its own challenges - and the new license is the cherry on the cake: it is locking you down to your own infrastructure unless you want to pay for timescale own SAAS offering (and then prey they do not alter the condition of the deal too much later)

It is just not worth it, unless you have a very small problem, or you can afford to have people concentrating on timescale maintenance - and in this case, you would be getting better bang for bucks by having these people work on clickhouse.

I'm speaking only from my own experience. I have relatively large servers dedicated to time series (about 100T of disk space, between 128 and 256 Gb of RAM). They were going to be retired for even bigger servers. Instead, we experimented with clickhouse on one of the recently decommissioned servers. We could not believe the benchmarks! Moving to clickhouse has improved the performance on about every metric. Yes, it required some minor SQL rewrites, about 1 day of work total, but unless your hardware is free and your queries are set in stone, clickhouse makes more sense.

mfreed 2153 days ago

Hi - I just want to clarify some mistake / misrepresentation about our Timescale License (TSL):

1. The TSL is not a new license, have had it in place since late 2018. What we recently announced is that multi-node TimescaleDB will be available for free under the TSL (free, source available), while (for example) clustered InfluxDB is purely proprietary (paid, closed source).

2. Our TSL license prevents offering TimescaleDB-as-a-service, it absolutely does NOT prevent you from running/offering a SaaS service or from utilizing cloud services/infra (you say "it is locking you down to your own infrastructure unless you want to pay for timescale own SAAS offering"). Specifically, Timescale offers a pure Apache-2 version and a "Timescale License" (TSL) Community version. For the TSL version, what it primarily restricts is the cloud providers like AWS and Azure from offering TimescaleDB-as-a-service (e.g., TimescaleDB Community on AWS RDS). Many thousands of companies use our community version for free to build SaaS services running on their own AWS instances.

> InfluxDB is purely proprietary (paid, closed source).

And clickhouse is not. I just suggest skipping the timescaledb step to someone migrating from influx, and going straight to clickhouse.

> For the TSL version, what it primarily restricts is the cloud providers like AWS and Azure from offering TimescaleDB-as-a-service (e.g., TimescaleDB Community on AWS RDS)

If there is some kind of emergency and I need to have the database on the cloud, this is a serious restriction. It limits my choices and constrains my actions.

> Many thousands of companies use our community version for free to build SaaS services running on their own AWS instances.

We have our servers, so it wasn't an issue. It was more of a long term concern, a chilling effect: what else may be restricted in the future?

Again, I think timescaledb has a wonderful place. It will certainly become the entry level database for timeseries.

It is just not suite for our workload.

[1] https://github.com/timescale/timescaledb-kubernetes

mfreed 2153 days ago

Hey, thanks for the continued discussion:

  > If there is some kind of emergency and I need to have
  > the database on the cloud, this is a serious restriction.
  > It limits my choices and constrains my actions.

  > if the license change to allow me to deploy to a cluster
  > of AWS servers (just in case we ditch our own hardawre),

You can deploy TimescaleDB on AWS servers (the TSL certainly allows it). Most of our users do. They don't run their own hardware.

You can even use our Apache-2 k8s helm charts [1] to immediately spin up a cluster of replicated TimescaleDB nodes with automated leader-election/failover and continuous incremental backup. The helm charts have first-class support for AWS EKS.

What the TSL prevents is _Amazon_ offering TimescaleDB as a paid DBaaS service. To my knowledge, none of the major cloud vendors offer Clickhouse as a first-class paid service, so that's somewhat a moot point. I guess theoretically Amazon could launch Clickhouse-as-a-service, but that theoretical possibility doesn't help you in your emergency.

fiddlerwoaroof 2153 days ago

Apart from license issues, Clickhouse is just really impressive: it’s a minor pain operationally, but in our tests it left all the Postgres-based timeseries solutions in the dust for real-time analytics without rollup tables.

manigandham 2153 days ago

Timescale has added their own layer of compression and columnar layouts to the Postgres row storage. That will get you to around 70% of the performance of using a dedicated column-oriented data warehouse, with the rest depending on how complex and selective your queries are.

It won't match the pure scan and computation speed of Clickhouse but the continuous aggregation feature is the recommended approach for querying large datasets (similar to Clickhouse table engines like AggregatingMergeTree).

[1] https://blog.timescale.com/blog/building-columnar-compressio...

cevian 2153 days ago

(TimescaleDB engineer here) Some of the comments here sound technically off.

We've never seen a backup take down a machine. The backups we use are the same as Postgres which are used by millions of companies without a problem (and can be streaming incremental backups like pgBackrest, WAL-E, etc. or whole-database backups like pg_dump). As with any DB you do have to size and configure your database correctly (which these days isn't hard).

We've never seen anybody claim that ClickHouse offers significantly better compression than we do overall. Obviously compression depends heavily on data distribution and I'm sure you could make up a dataset where clickhouse does better (just as you could where TimescaleDB does better). But on real distributions we don't see this at all, we do pretty advanced columnar compression on a per-datatype basis [1], and see median space reduction of 95% from compression across users.

Large queries is a weird claim to make since Postgres has more different types of indexes than Clickhouse and has support for multiple indexes. If you are processing all of your data for all your queries then yes, click house sequential scans may be better. But that's less common, and also where TimescaleDB continuous aggregates come in.

We've seen customers successfully use our single-node version with 100s of billions of rows so claiming that we are just for small use-cases is simply untrue, and especially with the launch of multi-node TimescaleDB.

I understand people may have different preferences and experiences, but some of these felt a bit off to me.

On servers with a very high CPU load, backup done without using continuous streaming to a second server and done from this second server, something (I have stopped using timescale so I can't tell you what did) during backup caused a peak in load and IO, impacting read and write performance of the primary server, causing a cascading failure of the processes due to timeouts, eventually taking the server down due to swap issues and OOM triggering a reboot.

So we stopped doing backups. Actually, that's how we started using clickhouse: for cold storage, as the files in /var/lib/clickhouse used far less storage space and issues. Eventually the same data was sent both to timescaledb and clickhouse, in a poor's man backup. Finally, timescaledb was removed.

> As with any DB you do have to size and configure your database correctly (which these days isn't hard).

Thanks for supposing we didn't try. We did not end up with 256Gb of RAM per server for no reason.

All I'm saying is that Timescale totally has a place, but not beyond a certain scale and complexity.

> We've never seen anybody claim that ClickHouse offers significantly better compression than we do overall

Altiny does, so do a few others. mandigandham above says that you are now at 70% of what clickhouse does. I'm not saying you're not improving. It was just one of the too many issues we had to fight.

Also, you have only recently introduced compression - good, but I'm not aware if you already offer something like DateTime Codec(DoubleDelta, LZ4), or the choice of compression algorithms. LZ4 can be slow, so there is a choice between various alternatives.

For example, T64 calculates the max and min values for the encoded range, and then strips the higher bits by transposing a 64-bit matrix. Sometimes it makes sense. zStd is slower than T64 but needs to scan less data, which makes up for it. Sometimes it makes more sense.

Large databases need more flexibility.

> If you are processing all of your data for all your queries then yes, click house sequential scans may be better

I confirm, it is better.

And for some workloads, continuous aggregates make no sense.

> We've seen customers successfully use our single-node version with 100s of billions of rows so claiming that we are just for small use-cases is simply untrue, and especially with the launch of multi-node TimescaleDB

I have about 50Tb of data per server. What is below 1Tb I call "small use cases".

> I understand people may have different preferences and experiences, but some of these felt a bit off to me.

When I was trying to use timescaledb and reported weird issues, I had the same return: my use case and bug report felt "off" to the person I reported them to.

Maybe it is why they weren't addressed - or maybe much later, when reported by more clients?

Personally, I have no horse in the game. If you become better than clickhouse for my workload, and if the license change to allow me to deploy to a cluster of AWS servers (just in case we ditch our own hardawre), I will consider timescale again in the future.

For now, I'm watching it evolve, and slowly address the outstanding issues, like disk usage, and performance. By your own admission and benchmarks, you are now at 70% of what clickhouse does - in my experience, the actual difference is much higher.

But I sincerely hope you succeed and catch up, as more software diversity is always better.

[0] https://www.altinity.com/blog/clickhouse-for-time-series

akulkarni 2153 days ago

Thanks for the thoughtful response.

On the compression point:

- I believe the Altinity Benchmarks [0] are from 2018, on TimescaleDB 0.12.1. TimescaleDB has gotten much better since then (now on version 1.7.2), and most notably, offers native compression now (it did not then).

- I believe manigandham's 70% comment is more of an offhand estimate and not a concrete benchmark. But perhaps he can weigh in. :-)

- Re: compression algorithms, TimescaleDB now employs several best-in-class algorithms, including delta-delta, gorilla, Simple-8b RLE. Much more than just LZ4. [1]

Overall, I don't think anyone has done a real storage comparison between TimescaleDB and Clickhouse since we launched native compression. It's on our todo list, but we also welcome external benchmarks. But based on what we've found versus similar systems, I suspect our storage usage would be really similar.

[1] https://blog.timescale.com/blog/time-series-compression-algo...

samhld 2153 days ago

(InfluxData solution architect here) Boolean is supported. You query it in the WHERE clause. Try `SELECT MAX(temperature) FROM ... WHERE temperature > 10`. That said, I'm not sure why you'd run a query like that in InfluxQL as it's the same as `SELECT max(temperature)`. :).

RedShift1 2153 days ago

It's not the same thing. 'SELECT MAX(temperature) > 10 FROM ...' gets you all the datapoints, with a value of true or false. Moving it to the WHERE clause only gets you the datapoints where temperature is > 10. Yes, you can fill with 0 after the GROUP BY, but if all datapoints are less than 10, you get nothing back from the database. That's confusing: were all datapoints less than 10, or was there nothing registered in the time window I'm querying? Impossible to tell the difference. Not to mention some user interfaces just bailing: no data. I need to show a chart with a timeline of this condition being true or false.

Plus my actual use case is even more complex, not only do I need something like MAX(temperature) > 10, I need (MAX(temperature)) > 10 && (MAX(temperature) - MAX(dewpoint)) > 4.5).

samhld 2153 days ago

Yep you're right -- oversight on my part. To be clear, the "not sure why you'd run a query like that" was referring to doing it the InfluxQL way (which is not the same thing)...where your results would end up being the same controlling for time range.

willvarfar 2153 days ago

Question for influx solution architect:

How do you delete points from a specific measurement in a specific retention policy?

not a solution architect but, just delete by a very specific timestamp. this is not possible if you are writing at coarse time precisions, so don't do that.

`delete from meas_name where time='2020-07-22T15:40:58.375506762Z'` - just tried that and it deleted one row.

willvarfar 2152 days ago

You didn’t specify the retention policy...?

kev009 2153 days ago

Blame the customer. Nice solution, architect!

samhld 2153 days ago

Huh? I don't mean to come off as blaming anyone for anything. Trying to offer a work around for their problem with InfluxQL. They usually exist. They may not be good enough for some but offering them doesn't hurt. In this case, I certainly misunderstood the need though.

it's not that hard: `select max from (select max(temperature) from measurement_name group by time(1h)) where max > 10`

RedShift1 2152 days ago

Does not produce the desired result, see elsewhere in this thread for why this doesn't work

ah, I see what you wanted now (`max(col) > 10...` -> [true, false, true, ..]). seems like it would be pretty easy to use numerical transformations for the same thing, depending on how you were going to be using it: `select max(col) - 10 from meas group by time(1h)` will give you negative results wherever you would have "false" in the query you wanted to do. primarily I want the timeseries/metrics database I use to be FAST over featureful, it's a specialized tool, and of course you can always just pull the data out and do whatever you want with it in code. which is to say, maybe you would be better off with something else, but that isn't necessarily a knock on influx, you know?

RedShift1 2152 days ago

Depends on the use case. If you only need to do some basic stuff like give me room temperature for that room between t1 and t2, or average server usage for all servers, fine. But any more business or analytical queries like "timeline of when was this true or false", "histogram of temperatures", "Power consumption this month", "runtime in hours per shift (where each shift is an arbitrary division in time)" are left in the cold with seemingly little progress made to fit these use cases. So for a system that claims real-time visibility it has suprisingly little analytical and time-based capabilities.