My Activities<\/h2>\n\nNOTE:<\/strong> I translated every talk title; the original\ntitles are in PT-BR, so some details may have been lost in\ntranslation.<\/p>\n<\/blockquote>\nI presented three talks and led one BoF session. The talks are all\navailable on Debian's Peertube:<\/p>\n
\nGet to\nknow dh-make-vim<\/a><\/li>\nZero-Code\nInstrumentation of an application using eBPF<\/a><\/li>\nDD\nnon-uploading can upload<\/a><\/li>\n<\/ul>\nYou can also find my slides at people.d.o<\/a>.<\/p>\nMy first talk was a showcase of dh-make-vim<\/a>, a\ntool I created and have been using for a few months. Some people tested\nit and found bugs, which was really nice to see.<\/p>\nMy second talk was essentially a live version of my blog post Zero-Code\nInstrumentation of an Envoy TCP Proxy using eBPF<\/a>.<\/p>\nI also gave a lightning talk about something many people are not\naware of: non-uploading DDs can also sponsor uploads.<\/p>\n
If you're interested, this bug report provides more context: tracker.debian.org:\nSigned by field is missing when sponsoring as DD non-uploading<\/a><\/p>\nFinally, I led the BoF session \"Experiences,\nlessons learned, and next steps from the mentoring sessions\"<\/a>. This\nwas my favorite session, we had many participants with different\nperspectives and ideas, which led to a very engaging discussion. I'm\nstill working on the action plans and I plan to release them soon.<\/p>\nHere are some photos of these activities:<\/p>\n
![](\".\/assets\/minidebconf-campinas-2026\/bof_1.jpeg\"\nalt=\"Mentorship)
<\/p>\n
![](\".\/assets\/minidebconf-campinas-2026\/bof_2.jpeg\"\nalt=\"Mentorship)
<\/p>\n
![](\".\/assets\/minidebconf-campinas-2026\/dd-non-uploader-uploading.jpeg\"\nalt=\"DD)
<\/p>\n
![](\".\/assets\/minidebconf-campinas-2026\/dh-make-vim.jpeg\"\nalt=\"dh-make-vim)
<\/p>\n
![](\"assets\/minidebconf-campinas-2026\/obi.jpeg\"\nalt=\"Zero-Code)
<\/p>\n
My favorite activities<\/h2>\n
This is a list, in no particular order, of some of the sessions I\nenjoyed the most:<\/p>\n
\nSalsa\nCI, showing features that almost nobody knows<\/a><\/p>\nI wrote a blog post<\/a>\nabout one of the things I learned in this talk, and there is still a lot\nmore to explore. Aquila Macedo is developing many cool features in Salsa\nCI.<\/p><\/li>\nFree\nSoftware: Freedom, Autonomy, Sovereignty<\/a><\/p>\nI had been really looking forward to this one. Alexandre Oliva is a\nvery important figure in the Free Software movement, especially in South\nAmerica. I'll need to rewatch it, my futures talks about Free Software\nwill likely be inspired by this one.<\/p><\/li>\n
What\nI've lived\/seen in 33 Years of Debian & Free Software in\ngeneral<\/a><\/p>\nEduardo Ma\u00e7an was the first Debian Developer in Brazil, so it's\nalways valuable to hear the story from someone who was part of\nit.<\/p><\/li>\n
Symbolism\n- an introduction<\/a><\/p>\nDespite the title, this talk was not about astrology! I'll probably\nrewatch it as well, as there is a lot of information to take in. I\nreally like the passion S\u00e9rgio Durigan has for C. He is also a great\nspeaker and knows how to guide the audience through the topic.<\/p><\/li>\n
Debate:\nContemporary controversies in Debian<\/a><\/p>\nThe debate itself was great, but the conversations we had afterward\nwere even better. I changed some of my opinions after hearing different\nperspectives. I don't think this format would work at DebConf, but I\nwould definitely like to attend another one like this.<\/p><\/li>\n
Why\nLTS on Debian?<\/a><\/p>\nI had a few questions about LTS, and Kanashiro and Santiago answered\nthem both during the talk and in the Q&A session. They also shared\nsome challenges and how to avoid them, it was a great learning\nexperience.<\/p><\/li>\n
From\nmy first contribution to the Debian Maintainer<\/a><\/p>\nPolkorny was a bit shy but did a great job! I really enjoy this kind\nof talk. It is always nice to see the different paths people\ntake.<\/p><\/li>\n<\/ul>\n
Unfortunatly, I couldn't attend everything I was interested in, as\nalways.<\/p>\n
DayTrip - The\nBrazilian Particle Accelerator<\/h2>\n
Sirius is the largest and most complex scientific infrastructure ever\nbuilt in Brazil and one of the most advanced synchrotron light sources\nin the world. My jaw dropped the entire time; it's hard to describe how\nincredible this is.<\/p>\n
My favorite detail: they're running Debian :)<\/p>\n
![](\".\/assets\/minidebconf-campinas-2026\/daytrip-1.jpeg\"\nalt=\"\"\nclass=\"markdown-img\")
<\/p>\n
![](\".\/assets\/minidebconf-campinas-2026\/daytrip-2.jpeg\"\nalt=\"\"\nclass=\"markdown-img\")
<\/p>\n
![](\".\/assets\/minidebconf-campinas-2026\/daytrip-3.jpeg\"\nalt=\"\"\nclass=\"markdown-img\")
<\/p>\n
![](\".\/assets\/minidebconf-campinas-2026\/daytrip-4.jpeg\"\nalt=\"\"\nclass=\"markdown-img\")
<\/p>\n
![](\".\/assets\/minidebconf-campinas-2026\/daytrip-5.jpeg\"\nalt=\"\"\nclass=\"markdown-img\")
<\/p>\n
![](\".\/assets\/minidebconf-campinas-2026\/daytrip-6.jpeg\"\nalt=\"\"\nclass=\"markdown-img\")
<\/p>\n
Wrap up<\/h2>\n
I believe this was the best MiniDebConf Brazil so far. There were\nmany other things I chose not to include here, as this post is already\nquite long. Still, here are a few more highlights:<\/p>\n
\n- A Bug Squashing Party<\/li>\n
- Driving Samuel Henrique's drones<\/li>\n
- Lots of capybaras<\/li>\n
- A small birthday party<\/li>\n
- A visit to two data centers<\/li>\n<\/ul>","link":{"@attributes":{"href":"minidebconf-campinas-2026.html"}},"id":"tag:www.sergiocipriano.com,2026-04-30:posts\/minidebconf-campinas-2026.md","published":"2026-04-30T16:14:47-03:00","updated":"2026-05-13T10:05:21-03:00"},{"title":"How to build reverse dependencies using Salsa CI","content":"
How to build\nreverse dependencies using Salsa CI<\/h1>\n
Last week, I attended MiniDebConf Campinas, and one of my favorites\ntalks was \"Salsa CI,\nshowing features that almost nobody knows\"<\/a> by Aquila Macedo.<\/p>\nOne of the things I learned is that we can easily build reverse\ndependencies using:<\/p>\n
$ git push -o ci.variable="SALSA_CI_DISABLE_BUILD_REVERSE_DEPENDENCIES=0"<\/code><\/pre>\nI tried this option before uploading typer version 0.20.0-1:<\/p>\n
![\"example](\".\/assets\/salsaci-build-rdeps\/example.png\")
<\/p>\n
This is an amazing feature. Thanks to everyone involved in making it\nhappen!<\/p>","link":{"@attributes":{"href":"salsaci-build-rdeps.html"}},"id":"tag:www.sergiocipriano.com,2026-04-29:posts\/salsaci-build-rdeps.md","published":"2026-04-29T23:27:46-03:00","updated":"2026-04-29T23:27:46-03:00"},{"title":"How to view the Debian Upload Queue","content":"
How to view the Debian\nUpload Queue<\/h1>\n
Some people may not know this, but the Debian Upload Queue is public\nand very easy to access:<\/p>\n
$ curl ftp:\/\/ftp.upload.debian.org\/pub\/UploadQueue\/\ndrwxr-sr-x 18 1518 1281 4096 Jun 26 2019 DELAYED\n-rw-r--r-- 1 1518 1281 3442 Jul 14 2025 README\n-rw-r----- 1 117 1281 3052 Apr 20 21:32 neovim-tokyonight_4.14.1-1.debian.tar.xz\n-rw-r----- 1 117 1281 2119 Apr 20 21:32 neovim-tokyonight_4.14.1-1.dsc\n-rw-r----- 1 117 1281 5533 Apr 20 21:32 neovim-tokyonight_4.14.1-1_amd64.buildinfo\n-rw-r----- 1 117 1281 2637 Apr 20 21:32 neovim-tokyonight_4.14.1-1_source.changes\n-rw-r----- 1 117 1281 197584 Apr 20 21:32 neovim-tokyonight_4.14.1.orig.tar.gz<\/code><\/pre>","link":{"@attributes":{"href":"list-upload-queue.html"}},"id":"tag:www.sergiocipriano.com,2026-04-21:posts\/list-upload-queue.md","published":"2026-04-21T12:16:58-03:00","updated":"2026-04-21T12:16:58-03:00"},{"title":"Zero-Code Instrumentation of an Envoy TCP Proxy using eBPF","content":"Zero-Code\nInstrumentation of an Envoy TCP Proxy using eBPF<\/h1>\nI recently had to debug an Envoy\nNetwork Load Balancer<\/a>, and the options Envoy provides just weren't\nenough. We were seeing a small number of HTTP 499 errors caused by\nlatency somewhere in our cloud<\/a>, but\nit wasn't clear what the bottleneck was. As a result, each team had to\nset up additional instrumentation to catch latency spikes and figure out\nwhat was going on.<\/p>\nMy team is responsible for the LBaaS product (Load Balancer as a\nService) and, of course, we are the first suspects when this kind of\nproblem appear.<\/p>\n
Before going for the current solution, I read a lot of Envoy's\ndocumentation.<\/p>\n
It is possible to enable access\nlogs<\/a> for Envoy, but they don't provide the information required for\nthis debug. This is an example of the output:<\/p>\n[2025-12-08T20:44:49.918Z] "- - -" 0 - 78 223 1 - "-" "-" "-" "-" "172.18.0.2:8080"<\/code><\/pre>\nI won't go into detail about the line above, since it's not possible\nto trace the request using access logs alone.<\/p>\n
Envoy also has OpenTelemetry\ntracing<\/a>, which is perfect for understanding sources of latency.\nUnfortunatly, it is only available for Application Load Balancers.<\/p>\nMost of the HTTP 499 were happening every 10 minutes, so we managed\nto get some of the requests with tcpdump, Wireshark<\/a> and using http headers\nto filter the requests.<\/p>\nThis approach helped us reproduce and track down the problem, but it\nwasn't a great solution. We clearly needed better tools to catch this\nkind of issue the next time it happened.<\/p>\n
Therefore, I decided to try out OpenTelemetry\neBPF Instrumentation<\/a>, also referred to as OBI.<\/p>\nI saw the announcement of Grafana Beyla before it was renamed to OBI,\nbut I didn't have the time or a strong reason to try it out until now.\nEven then, I really liked the idea, and the possibility of using eBPF to\nsolve this instrumentation problem had been in the back of my mind.<\/p>\n
OBI promises zero-code automatic instrumentation for Linux services\nusing eBPF, so I put together a minimal setup to see how well it\nworks.<\/p>\n
Reproducible setup<\/h2>\n
I used the following tools:<\/p>\n
\n- Docker Compose<\/a><\/li>\n
- EnvoyProxy<\/a><\/li>\n
- Go<\/a><\/li>\n
OBI<\/a><\/li>\n<\/ul>\nSetting up a TCP Proxy with Envoy was straightforward:<\/p>\n
<\/a>static_resources<\/span>:<\/span><\/span>\n<\/a> <\/span>listeners<\/span>:<\/span><\/span>\n<\/a> <\/span>-<\/span> <\/span>name<\/span>:<\/span> go_server_listener<\/span><\/span>\n<\/a> <\/span>address<\/span>:<\/span><\/span>\n<\/a> <\/span>socket_address<\/span>:<\/span><\/span>\n<\/a> <\/span>address<\/span>:<\/span> <\/span>0.0.0.0<\/span><\/span>\n<\/a> <\/span>port_value<\/span>:<\/span> <\/span>8000<\/span><\/span>\n<\/a> <\/span>filter_chains<\/span>:<\/span><\/span>\n<\/a> <\/span>-<\/span> <\/span>filters<\/span>:<\/span><\/span>\n<\/a> <\/span>-<\/span> <\/span>name<\/span>:<\/span> envoy.filters.network.tcp_proxy<\/span><\/span>\n<\/a> <\/span>typed_config<\/span>:<\/span><\/span>\n<\/a> <\/span>"@type"<\/span>:<\/span> type.googleapis.com\/envoy.extensions.filters.network.tcp_proxy.v3.TcpProxy<\/span><\/span>\n<\/a> <\/span>stat_prefix<\/span>:<\/span> go_server_tcp<\/span><\/span>\n<\/a> <\/span>cluster<\/span>:<\/span> go_server_cluster<\/span><\/span>\n<\/a> <\/span>clusters<\/span>:<\/span><\/span>\n<\/a> <\/span>-<\/span> <\/span>name<\/span>:<\/span> go_server_cluster<\/span><\/span>\n<\/a> <\/span>connect_timeout<\/span>:<\/span> 1s<\/span><\/span>\n<\/a> <\/span>type<\/span>:<\/span> LOGICAL_DNS<\/span><\/span>\n<\/a> <\/span>load_assignment<\/span>:<\/span><\/span>\n<\/a> <\/span>cluster_name<\/span>:<\/span> go_server_cluster<\/span><\/span>\n<\/a> <\/span>endpoints<\/span>:<\/span><\/span>\n<\/a> <\/span>-<\/span> <\/span>lb_endpoints<\/span>:<\/span><\/span>\n<\/a> <\/span>-<\/span> <\/span>endpoint<\/span>:<\/span><\/span>\n<\/a> <\/span>address<\/span>:<\/span><\/span>\n<\/a> <\/span>socket_address<\/span>:<\/span><\/span>\n<\/a> <\/span>address<\/span>:<\/span> target-backend<\/span><\/span>\n<\/a> <\/span>port_value<\/span>:<\/span> <\/span>8080<\/span><\/span><\/code><\/pre><\/div>\nThis is the simplest Envoy TCP proxy configuration: a listener on\nport 8000 forwarding traffic to a backend running on port 8080.<\/p>\n
For the backend, I used a basic Go HTTP server:<\/p>\n
NOTE:<\/strong> I translated every talk title; the original\ntitles are in PT-BR, so some details may have been lost in\ntranslation.<\/p>\n<\/blockquote>\n I presented three talks and led one BoF session. The talks are all\navailable on Debian's Peertube:<\/p>\n You can also find my slides at My first talk was a showcase of My second talk was essentially a live version of my blog post I also gave a lightning talk about something many people are not\naware of: non-uploading DDs can also sponsor uploads.<\/p>\n If you're interested, this bug report provides more context: Finally, I led the BoF session Here are some photos of these activities:<\/p>\n This is a list, in no particular order, of some of the sessions I\nenjoyed the most:<\/p>\n I wrote a I had been really looking forward to this one. Alexandre Oliva is a\nvery important figure in the Free Software movement, especially in South\nAmerica. I'll need to rewatch it, my futures talks about Free Software\nwill likely be inspired by this one.<\/p><\/li>\n Eduardo Ma\u00e7an was the first Debian Developer in Brazil, so it's\nalways valuable to hear the story from someone who was part of\nit.<\/p><\/li>\n Despite the title, this talk was not about astrology! I'll probably\nrewatch it as well, as there is a lot of information to take in. I\nreally like the passion S\u00e9rgio Durigan has for C. He is also a great\nspeaker and knows how to guide the audience through the topic.<\/p><\/li>\n The debate itself was great, but the conversations we had afterward\nwere even better. I changed some of my opinions after hearing different\nperspectives. I don't think this format would work at DebConf, but I\nwould definitely like to attend another one like this.<\/p><\/li>\n I had a few questions about LTS, and Kanashiro and Santiago answered\nthem both during the talk and in the Q&A session. They also shared\nsome challenges and how to avoid them, it was a great learning\nexperience.<\/p><\/li>\n Polkorny was a bit shy but did a great job! I really enjoy this kind\nof talk. It is always nice to see the different paths people\ntake.<\/p><\/li>\n<\/ul>\n Unfortunatly, I couldn't attend everything I was interested in, as\nalways.<\/p>\n Sirius is the largest and most complex scientific infrastructure ever\nbuilt in Brazil and one of the most advanced synchrotron light sources\nin the world. My jaw dropped the entire time; it's hard to describe how\nincredible this is.<\/p>\n My favorite detail: they're running Debian :)<\/p>\n I believe this was the best MiniDebConf Brazil so far. There were\nmany other things I chose not to include here, as this post is already\nquite long. Still, here are a few more highlights:<\/p>\n Last week, I attended MiniDebConf Campinas, and one of my favorites\ntalks was One of the things I learned is that we can easily build reverse\ndependencies using:<\/p>\n I tried this option before uploading typer version 0.20.0-1:<\/p>\n This is an amazing feature. Thanks to everyone involved in making it\nhappen!<\/p>","link":{"@attributes":{"href":"salsaci-build-rdeps.html"}},"id":"tag:www.sergiocipriano.com,2026-04-29:posts\/salsaci-build-rdeps.md","published":"2026-04-29T23:27:46-03:00","updated":"2026-04-29T23:27:46-03:00"},{"title":"How to view the Debian Upload Queue","content":" Some people may not know this, but the Debian Upload Queue is public\nand very easy to access:<\/p>\n I recently had to debug an My team is responsible for the LBaaS product (Load Balancer as a\nService) and, of course, we are the first suspects when this kind of\nproblem appear.<\/p>\n Before going for the current solution, I read a lot of Envoy's\ndocumentation.<\/p>\n It is possible to enable I won't go into detail about the line above, since it's not possible\nto trace the request using access logs alone.<\/p>\n Envoy also has Most of the HTTP 499 were happening every 10 minutes, so we managed\nto get some of the requests with tcpdump, This approach helped us reproduce and track down the problem, but it\nwasn't a great solution. We clearly needed better tools to catch this\nkind of issue the next time it happened.<\/p>\n Therefore, I decided to try out I saw the announcement of Grafana Beyla before it was renamed to OBI,\nbut I didn't have the time or a strong reason to try it out until now.\nEven then, I really liked the idea, and the possibility of using eBPF to\nsolve this instrumentation problem had been in the back of my mind.<\/p>\n OBI promises zero-code automatic instrumentation for Linux services\nusing eBPF, so I put together a minimal setup to see how well it\nworks.<\/p>\n I used the following tools:<\/p>\n Setting up a TCP Proxy with Envoy was straightforward:<\/p>\n This is the simplest Envoy TCP proxy configuration: a listener on\nport 8000 forwarding traffic to a backend running on port 8080.<\/p>\n For the backend, I used a basic Go HTTP server:<\/p>\n\n
<\/p>\n<\/p>\n<\/p>\n<\/p>\n<\/p>\nMy favorite activities<\/h2>\n
\n
DayTrip - The\nBrazilian Particle Accelerator<\/h2>\n
<\/p>\n<\/p>\n<\/p>\n<\/p>\n<\/p>\n<\/p>\nWrap up<\/h2>\n
\n
How to build\nreverse dependencies using Salsa CI<\/h1>\n
$ git push -o ci.variable="SALSA_CI_DISABLE_BUILD_REVERSE_DEPENDENCIES=0"<\/code><\/pre>\n<\/p>\nHow to view the Debian\nUpload Queue<\/h1>\n
$ curl ftp:\/\/ftp.upload.debian.org\/pub\/UploadQueue\/\ndrwxr-sr-x 18 1518 1281 4096 Jun 26 2019 DELAYED\n-rw-r--r-- 1 1518 1281 3442 Jul 14 2025 README\n-rw-r----- 1 117 1281 3052 Apr 20 21:32 neovim-tokyonight_4.14.1-1.debian.tar.xz\n-rw-r----- 1 117 1281 2119 Apr 20 21:32 neovim-tokyonight_4.14.1-1.dsc\n-rw-r----- 1 117 1281 5533 Apr 20 21:32 neovim-tokyonight_4.14.1-1_amd64.buildinfo\n-rw-r----- 1 117 1281 2637 Apr 20 21:32 neovim-tokyonight_4.14.1-1_source.changes\n-rw-r----- 1 117 1281 197584 Apr 20 21:32 neovim-tokyonight_4.14.1.orig.tar.gz<\/code><\/pre>","link":{"@attributes":{"href":"list-upload-queue.html"}},"id":"tag:www.sergiocipriano.com,2026-04-21:posts\/list-upload-queue.md","published":"2026-04-21T12:16:58-03:00","updated":"2026-04-21T12:16:58-03:00"},{"title":"Zero-Code Instrumentation of an Envoy TCP Proxy using eBPF","content":"[2025-12-08T20:44:49.918Z] "- - -" 0 - 78 223 1 - "-" "-" "-" "-" "172.18.0.2:8080"<\/code><\/pre>\nReproducible setup<\/h2>\n
\n
<\/a>static_resources<\/span>:<\/span><\/span>\n<\/a> <\/span>listeners<\/span>:<\/span><\/span>\n<\/a> <\/span>-<\/span> <\/span>name<\/span>:<\/span> go_server_listener<\/span><\/span>\n<\/a> <\/span>address<\/span>:<\/span><\/span>\n<\/a> <\/span>socket_address<\/span>:<\/span><\/span>\n<\/a> <\/span>address<\/span>:<\/span> <\/span>0.0.0.0<\/span><\/span>\n<\/a> <\/span>port_value<\/span>:<\/span> <\/span>8000<\/span><\/span>\n<\/a> <\/span>filter_chains<\/span>:<\/span><\/span>\n<\/a> <\/span>-<\/span> <\/span>filters<\/span>:<\/span><\/span>\n<\/a> <\/span>-<\/span> <\/span>name<\/span>:<\/span> envoy.filters.network.tcp_proxy<\/span><\/span>\n<\/a> <\/span>typed_config<\/span>:<\/span><\/span>\n<\/a> <\/span>"@type"<\/span>:<\/span> type.googleapis.com\/envoy.extensions.filters.network.tcp_proxy.v3.TcpProxy<\/span><\/span>\n<\/a> <\/span>stat_prefix<\/span>:<\/span> go_server_tcp<\/span><\/span>\n<\/a> <\/span>cluster<\/span>:<\/span> go_server_cluster<\/span><\/span>\n<\/a> <\/span>clusters<\/span>:<\/span><\/span>\n<\/a> <\/span>-<\/span> <\/span>name<\/span>:<\/span> go_server_cluster<\/span><\/span>\n<\/a> <\/span>connect_timeout<\/span>:<\/span> 1s<\/span><\/span>\n<\/a> <\/span>type<\/span>:<\/span> LOGICAL_DNS<\/span><\/span>\n<\/a> <\/span>load_assignment<\/span>:<\/span><\/span>\n<\/a> <\/span>cluster_name<\/span>:<\/span> go_server_cluster<\/span><\/span>\n<\/a> <\/span>endpoints<\/span>:<\/span><\/span>\n<\/a> <\/span>-<\/span> <\/span>lb_endpoints<\/span>:<\/span><\/span>\n<\/a> <\/span>-<\/span> <\/span>endpoint<\/span>:<\/span><\/span>\n<\/a> <\/span>address<\/span>:<\/span><\/span>\n<\/a> <\/span>socket_address<\/span>:<\/span><\/span>\n<\/a> <\/span>address<\/span>:<\/span> target-backend<\/span><\/span>\n<\/a> <\/span>port_value<\/span>:<\/span> <\/span>8080<\/span><\/span><\/code><\/pre><\/div>\n