Metrics to help relay operators?
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi Alison, moving this thread from the metrics-team@ moderation queue to this list with your permission:
Hi Metrics team,
I was talking to Roger on IRC about how great it is that he sent those checkin mails to everyone who runs a fast relay, and he had an idea for the metrics and community teams to work together to get useful lists that we could use to help give better support to our network of relay operators. For example, a weekly mail of all the relays that just passed the "earned a tshirt" threshold, or some other threshold about how much bandwidth they've pushed over a period of time, etc. The community team could then contact the relay operator(s) and congratulate them, thank them for their service, see if we can offer them help in any other ways, and so forth.
I've copied the community team here. Let's chat if you guys are interested!
Alison
Sure, let's do this. I might not be able to hack much on this myself in the next weeks, but I'm happy to speculate how such a tool could be designed. Is there anybody in the community team (or another friendly volunteer on this list) who'd be comfortable writing a small Python script that fetches Onionoo data [1] and filters out contact information of relays matching given criteria? That's how Compass [2] was started before it became a website, and it's how the t-shirt script [3] came to life (shortly before it was forgotten). I could imagine that we start with a simple Python script that can be run whenever somebody from the community team is ready to reach out to new relay operators. We could later automate that process by letting cron run that script once per week and have it send the output to some mailing list. But that could be step two. Hope that helps as a start. All the best, Karsten [1] https://onionoo.torproject.org/ [2] https://compass.torproject.org/ [3] https://gitweb.torproject.org/metrics-tasks.git/tree/task-9889/tshirt.py -----BEGIN PGP SIGNATURE----- Comment: GPGTools - http://gpgtools.org iQEcBAEBAgAGBQJXxT3zAAoJEC3ESO/4X7XBilwH/21JJ6HS8Nr1gfIrOWGurAQl qJBSref2NoPsXJuwG1Z038Zq6JysTRNkzZoqgmZTbILwnMAoXM8rXRZuUQnCGgxN jP5V0xjtSM6d7VRLr34JXt8pbo1Cqd1/tGzEHzsS2muUA80vZK9klv/M6O5CS0wE nqjApwD2OHRBVPq4D+MX2DxV705DJTCgCF3caA8z2nocs0oKDzC8078Wy4usc4KB l9Yf/WLyv24Ny2Uq7m8C/WkyAYEOBAQwze9/Z7leTASUaF9EtSxJKbdx1dIUexgv tTCr7mEBRqdAeKfUNUmrQxBkJdZxLUAnHGjuJrmj+GHrXawmsLkguu4dNf23sKc= =i5Pp -----END PGP SIGNATURE-----
On 30 Aug (10:04:03), Karsten Loesing wrote:
Hi Alison,
moving this thread from the metrics-team@ moderation queue to this list with your permission:
Hi Metrics team,
I was talking to Roger on IRC about how great it is that he sent those checkin mails to everyone who runs a fast relay, and he had an idea for the metrics and community teams to work together to get useful lists that we could use to help give better support to our network of relay operators. For example, a weekly mail of all the relays that just passed the "earned a tshirt" threshold, or some other threshold about how much bandwidth they've pushed over a period of time, etc. The community team could then contact the relay operator(s) and congratulate them, thank them for their service, see if we can offer them help in any other ways, and so forth.
I've copied the community team here. Let's chat if you guys are interested!
Alison
Sure, let's do this.
I might not be able to hack much on this myself in the next weeks, but I'm happy to speculate how such a tool could be designed.
Is there anybody in the community team (or another friendly volunteer on this list) who'd be comfortable writing a small Python script that fetches Onionoo data [1] and filters out contact information of relays matching given criteria?
I do have lots of scripts that interacts with CollecTor and uses stem to parse it all. I've used Onionoo few times but with bash scripts and grep/awk magic. However, I'm familiar with it so using Python for this would be easier I believe. If someone can specify for me the criteria for some scripts, I'm happy to do it, it will take me few minutes I believe with what I have already.
That's how Compass [2] was started before it became a website, and it's how the t-shirt script [3] came to life (shortly before it was forgotten).
I could imagine that we start with a simple Python script that can be run whenever somebody from the community team is ready to reach out to new relay operators.
We could later automate that process by letting cron run that script once per week and have it send the output to some mailing list. But that could be step two.
I can even do that as well! I already have so many scripts and cron running to detect bad relays on our network and measure the network[1] so if I know what we want from someone from the Community team, easy peasy lemon squeezy. Cheers! David [1] http://ygzf7uqcusp4ayjs.onion/tor-health/tor-health/index.html
Hope that helps as a start.
All the best, Karsten
[1] https://onionoo.torproject.org/
[2] https://compass.torproject.org/
[3] https://gitweb.torproject.org/metrics-tasks.git/tree/task-9889/tshirt.py
_______________________________________________ tor-project mailing list tor-project@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-project
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 30/08/16 15:14, David Goulet wrote:
On 30 Aug (10:04:03), Karsten Loesing wrote:
Hi Alison,
moving this thread from the metrics-team@ moderation queue to this list with your permission:
Hi Metrics team,
I was talking to Roger on IRC about how great it is that he sent those checkin mails to everyone who runs a fast relay, and he had an idea for the metrics and community teams to work together to get useful lists that we could use to help give better support to our network of relay operators. For example, a weekly mail of all the relays that just passed the "earned a tshirt" threshold, or some other threshold about how much bandwidth they've pushed over a period of time, etc. The community team could then contact the relay operator(s) and congratulate them, thank them for their service, see if we can offer them help in any other ways, and so forth.
I've copied the community team here. Let's chat if you guys are interested!
Alison
Sure, let's do this.
I might not be able to hack much on this myself in the next weeks, but I'm happy to speculate how such a tool could be designed.
Is there anybody in the community team (or another friendly volunteer on this list) who'd be comfortable writing a small Python script that fetches Onionoo data [1] and filters out contact information of relays matching given criteria?
I do have lots of scripts that interacts with CollecTor and uses stem to parse it all. I've used Onionoo few times but with bash scripts and grep/awk magic. However, I'm familiar with it so using Python for this would be easier I believe.
I suggest we use Onionoo for this rather than CollecTor. Onionoo already keeps a lot of history about relays, and we'd have to re-implement that functionality when using CollecTor data. Maybe take a look at the t-shirt script that I included as [3] in my original posting (link still below). That script uses Onionoo data, and it should already be relatively close to what community folks want.
If someone can specify for me the criteria for some scripts, I'm happy to do it, it will take me few minutes I believe with what I have already.
How about you and I participate in one of the next community team meetings and discuss possible requirements with community team people? Alison, would that be helpful? All the best, Karsten
That's how Compass [2] was started before it became a website, and it's how the t-shirt script [3] came to life (shortly before it was forgotten).
I could imagine that we start with a simple Python script that can be run whenever somebody from the community team is ready to reach out to new relay operators.
We could later automate that process by letting cron run that script once per week and have it send the output to some mailing list. But that could be step two.
I can even do that as well! I already have so many scripts and cron running to detect bad relays on our network and measure the network[1] so if I know what we want from someone from the Community team, easy peasy lemon squeezy.
Cheers! David
[1] http://ygzf7uqcusp4ayjs.onion/tor-health/tor-health/index.html
Hope that helps as a start.
All the best, Karsten
[1] https://onionoo.torproject.org/
[2] https://compass.torproject.org/
[3] https://gitweb.torproject.org/metrics-tasks.git/tree/task-9889/tshirt.py
tor-project mailing list tor-project@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-project
tor-project mailing list tor-project@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-project -----BEGIN PGP SIGNATURE----- Comment: GPGTools - http://gpgtools.org
iQEcBAEBAgAGBQJXxo/jAAoJEC3ESO/4X7XBGe8H/1A+j2DVGy/QXzfjaSL7LqNF 80cxzi1xFqFIYs25iMf0MEOc7bzclmEOGOIj6LdxnKZLqtfQUrIlHxDDy5IPTAa/ /ySTnKybVeTK97udqEVOHC75LuYZoAlJjPxiEeau5t30XqYJO+9f/FXGiUnoDiLH 298lf3RdOfFq09joxBhYIXBtRoRcYX8ZrWUDNnFu/Q5qmH2lUOfgfHtDZnu7mBzL ttOYgOZ8+MGqunwFV4sTMNUX4X21KLrdoXucKgBr65/GCsfGZLunCvcULIMVM1wF DBVKC7h0WH5xAIwAiVMfWXZv2ddSo33+VPCvpqZSIodfPfd5i5AU6KwuHTyJYxA= =4nWL -----END PGP SIGNATURE-----
Karsten Loesing:
If someone can specify for me the criteria for some scripts, I'm happy to do it, it will take me few minutes I believe with what I have already.
How about you and I participate in one of the next community team meetings and discuss possible requirements with community team people?
Sure, that sounds good. Our next meeting is Wednesday September 7 at 1800 UTC in #tor-project. In the meantime, I will send a mail to the tor-relays list and see if we can find out what would be valuable criteria for them. Alison
On 31 Aug (10:05:55), Karsten Loesing wrote:
On 30/08/16 15:14, David Goulet wrote:
On 30 Aug (10:04:03), Karsten Loesing wrote:
Hi Alison,
moving this thread from the metrics-team@ moderation queue to this list with your permission:
Hi Metrics team,
I was talking to Roger on IRC about how great it is that he sent those checkin mails to everyone who runs a fast relay, and he had an idea for the metrics and community teams to work together to get useful lists that we could use to help give better support to our network of relay operators. For example, a weekly mail of all the relays that just passed the "earned a tshirt" threshold, or some other threshold about how much bandwidth they've pushed over a period of time, etc. The community team could then contact the relay operator(s) and congratulate them, thank them for their service, see if we can offer them help in any other ways, and so forth.
I've copied the community team here. Let's chat if you guys are interested!
Alison
Sure, let's do this.
I might not be able to hack much on this myself in the next weeks, but I'm happy to speculate how such a tool could be designed.
Is there anybody in the community team (or another friendly volunteer on this list) who'd be comfortable writing a small Python script that fetches Onionoo data [1] and filters out contact information of relays matching given criteria?
I do have lots of scripts that interacts with CollecTor and uses stem to parse it all. I've used Onionoo few times but with bash scripts and grep/awk magic. However, I'm familiar with it so using Python for this would be easier I believe.
I suggest we use Onionoo for this rather than CollecTor. Onionoo already keeps a lot of history about relays, and we'd have to re-implement that functionality when using CollecTor data.
Maybe take a look at the t-shirt script that I included as [3] in my original posting (link still below). That script uses Onionoo data, and it should already be relatively close to what community folks want.
If someone can specify for me the criteria for some scripts, I'm happy to do it, it will take me few minutes I believe with what I have already.
How about you and I participate in one of the next community team meetings and discuss possible requirements with community team people?
Sure. I can be there. Thanks! David
Alison, would that be helpful?
All the best, Karsten
That's how Compass [2] was started before it became a website, and it's how the t-shirt script [3] came to life (shortly before it was forgotten).
I could imagine that we start with a simple Python script that can be run whenever somebody from the community team is ready to reach out to new relay operators.
We could later automate that process by letting cron run that script once per week and have it send the output to some mailing list. But that could be step two.
I can even do that as well! I already have so many scripts and cron running to detect bad relays on our network and measure the network[1] so if I know what we want from someone from the Community team, easy peasy lemon squeezy.
Cheers! David
[1] http://ygzf7uqcusp4ayjs.onion/tor-health/tor-health/index.html
Hope that helps as a start.
All the best, Karsten
[1] https://onionoo.torproject.org/
[2] https://compass.torproject.org/
[3] https://gitweb.torproject.org/metrics-tasks.git/tree/task-9889/tshirt.py
tor-project mailing list tor-project@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-project
tor-project mailing list tor-project@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-project
participants (3)
-
Alison -
David Goulet -
Karsten Loesing