My personal blog
hauleth.dev
blog
Fix links and typos in article about Elixir perf
+12
-4
+12
-4
content/post/things-about-elixir-you-probably-will-never-need/index.md
+12
-4
content/post/things-about-elixir-you-probably-will-never-need/index.md
···
15
15
"beam",
16
16
"performance"
17
17
]
18
+
19
+
[[extra.thanks]]
20
+
name = "Angry Clippy (@ze.du on Discord)"
21
+
why = "Redaction"
18
22
+++
19
23
20
24
In my last larger gig I worked on fascinating project - [Postgres connection
···
37
41
I can. This project now lives as [Ultravisor][] - it is still nowhere near being
38
42
done in a way that I like, but I still go back to work on it from time to time
39
43
to find potential performance improvements.
44
+
45
+
[Ultravisor]: https://github.com/Ultravisor/ultravisor
40
46
41
47
This is a story of things that I have done and learned during that journey.
42
48
···
74
80
times and then running `cat *.bggg` to concatenate all files into larger trace.
75
81
That has disadvantages, but at least it was workable within [Speedoscope][]
76
82
which I also highly recommend to anyone who needs to work on such optimisation.
83
+
84
+
[Speedoscope]: https://speedoscope.app
77
85
78
86
While flame graphs are awesome, there is cost to gathering them with eFlambè -
79
87
it greatly affects performance. Fortunately Erlang has some built in tools that
···
171
179
[Telemetry]: https://github.com/beam-telemetry/telemetry
172
180
173
181
In this project the metrics are exposed in Prometheus/OpenMetrics format, which
174
-
mean that there need to be collection system within application. In BEAM
182
+
means that there needs to be collection system within the application. In BEAM
175
183
applications the standard way to implement that is to use ETS tables to store
176
184
recorded values. Fortunately there are libraries to handle that for you, and for
177
185
the longest time "gold standard" for it was `telemetry_prometheus_core` library
···
285
293
286
294
## Lesson: Calling your `GenServer`s is fast, but not 90k times per second fast
287
295
288
-
One of the interesting observations is that I have spotted is that if there are
289
-
longer running queries, one that send more data over the network than just
290
-
simple short response, then the difference between Ultravisor and "state of the
296
+
One of the interesting observations that I have spotted is that if there are
297
+
longer running queries, ones that send more data over the network than just
298
+
simple short responses, then the difference between Ultravisor and "state of the
291
299
art" tools like [PgBouncer][] or [PgDog][] (that are written in non-managed
292
300
languages like C and Rust) is much smaller (obviously it is still there, but it
293
301
is on par, not substantially off).