Add mtr #127

jimaek · 2022-05-16T12:13:13Z

Example binary command:

mtr -4 -o "LDRAVM" --aslookup --show-ips --interval 0.5 --gracetime 3 --max-ttl 30 --timeout 15 -c 3 google.com
mtr -4 -o "LDRAVM" --aslookup --show-ips --interval 0.5 --gracetime 3 --max-ttl 30 --timeout 15 {--udp|--tcp} -c $PACKETS google.com

Output is a bit different from traceroute. It opens a new window basically and then exits. So to have real-time results we need to read the new window.
We also need to remove the first 3 lines of output (if it didnt fail with an error)
Not sure how to detect it finished. Maybe check for an exit code, it seems to return 0 when it finished without errors? echo $?

packets {6} - The number of packets to send. max 16. Default is 3 (optional)
protocol {TCP} - Mtr protocol. ICMP|TCP|UDP. Default is ICMP (optional)
port {90} - Port to use if TCP or UDP is selected above. Default 80 (optional)

The text was updated successfully, but these errors were encountered:

patrykcieszkowski · 2022-05-16T15:45:07Z

I'm not sure if that's gonna be a problem, but seems worth mentioning:

mtr: non-root users cannot request an interval < 1.0 seconds

jimaek · 2022-05-16T15:47:41Z

Yeah, it should be fine inside our containers :) But let me know if something unexpected comes up

patrykcieszkowski · 2022-05-16T16:06:20Z

another note:
flag -r runs MTR in stdout

jimaek · 2022-05-16T16:09:19Z

It also makes it not-real time :(

patrykcieszkowski · 2022-05-24T19:27:30Z

What should we do with MTR's ASCI color characters? They polute the output, to the point it makes it unreadable. Naturally, we gotta get rid of them for the parsed output, but should we keep them for the rawOutput?

{
  rawOutput: '\x1B(B\x1B)0\x1B[?1049h\x1B[1;24r\x1B[m\x0F\x1B[4l\x1B[39;49m\x1B[39;49m\x1B[m\x0F\x1B[H\x1B[J\x1B[1;29H\x1B[0;1m\x0F\x1B[1K My traceroute  [v0.93]\r\n' +
    '\x1B[m\x0Fdev-serv (192.168.0.16)\x1B[2;56H2022-05-24T19:21:42+0000\r\n' +
    'Keys:  \x1B[0;1m\x0FH\x1B[m\x0Felp   \x1B[0;1m\x0FD\x1B[m\x0Fisplay mode   \x1B[0;1m\x0FR\x1B[m\x0Festart statistics   \x1B[0;1m\x0FO\x1B[m\x0Frder of fields   \x1B[0;1m\x0Fq\x1B[m\x0Fuit\x1B[4;46H\x1B[0;1m\x0F   Packets               Pings\r\n' +
    ' Host\x1B[5;46H Loss% Drop   Rcv   Avg StDev Javg\r\x1BM\x1BM\x1B[m\x0F\n' +
    '\n' +
    '\n' +
    ' 1. AS???    192.168.0.1 (192.168.0.1)\x1B[6;48H0.0%    0     1   2.1   0.0  0.0\r\n' +
    ' 2. \x1B[0;1m\x0F(waiting for reply)\x1B[H\n' +
    '\n' +
    '\x1B[m\x0F\x1B[8d 3. AS5089   62.252.67.181 (62.252.67.181)     0.0%    0     1  18.4   0.0  0.0\r\n' +
    ' 4. \x1B[0;1m\x0F(waiting for reply)\x1B[H\n' +
    '\n' +
    '\x1B[m\x0F\x1B[8;14Hlutn-core-2a-xe-7114-0.network.v\x1B[H\n' +
    '\n' +
    '\x1B[10d 5. AS5089   62.254.59.130 (62.254.59.130)     0.0%    0     1  10.9   0.0  0.0\r\n' +
    ' 6. \x1B[0;1m\x0F(waiting for reply)\x1B[H\n' +
    '\n' +
    '\x1B[m\x0F\x1B[10;14Heislou2-ic-4-ae0-0.network.virgi\x1B[H\n' +
    '\n' +
    '\x1B[11;5HAS15169  142.250.160.116 (142.250.160.116  0.0%    0     1  12.7   0.0  0.0\r\n' +
    ' 7. \x1B[0;1m\x0F(waiting for reply)\x1B[H\n' +
    '\n' +
    '\x1B[m\x0F\x1B[12;5HAS15169  216.239.41.149 (216.239.41.149)   0.0%    0     1  14.5   0.0  0.0\r\n' +
    ' 8. \x1B[0;1m\x0F(waiting for reply)\x1B[H\n' +
    '\n' +
    '\x1B[m\x0F\x1B[13;5HAS15169  142.251.52.149 (142.251.52.149)   0.0%    0     1  26.4   0.0  0.0\r\n' +
    ' 9. \x1B[0;1m\x0F(waiting for reply)\x1B[H\n' +
    '\n' +
    '\x1B[m\x0F\x1BM\x1B[74G3\x1B[14;5HAS15169  172.217.16.238 (172.217.16.238)   0.0%    0     1  10.4   0.0  0.0\x1B[H\n' +
    '\n' +
    '\x1B[14;14Hmad08s04-in-f14.1e100.net (172.2\x1B[H\n' +
    '\n' +
    '\x1BM\x1B[74G4\r\n' +
    '\x1BM\x1B[74G5\r\n' +
    '\x1BM\x1B[74G6\r\n' +
    '\x1B[24;1H\x1B[?1049l\r\x1B[?1l\x1B>'
}

jimaek · 2022-05-24T19:31:58Z

definitely no, the question is if we need to do this manually or there is an option to disable it

patrykcieszkowski · 2022-05-24T19:32:52Z

I can easily filter them out with single regex query, that's not an issue. I just wasn't sure if we want to manipulate the rawOutput at all.

jimaek · 2022-05-24T19:35:55Z

There is no value in keeping the colors and it will make the life easier for many people if its regular readable text. So lets remove it :)

jimaek · 2022-06-14T17:32:31Z

Most rows return (waiting for reply) instead of the result. While the mtr command has them every time
ASNs are sometimes missing, but not via mtr
Last line is sometimes duplicated
Metadata should round the numbers "jMin": 0.15600000000000003,

patrykcieszkowski · 2022-06-15T00:17:02Z

Most rows return (waiting for reply) instead of the result. While the mtr command has them every time

It happened randomly, and was especially noticible on production probes - it didn't happen on my local dev serv. If I'm right, it was caused by the fact we have to asynchroniously query ASN values, while we tried to keep progress events going, so we ended up overriding the result. Fixed.

ASNs are sometimes missing, but not via mtr

fixed.

Last line is sometimes duplicated

True, but not true. Some routes point to the same server, and MTR responds with... new routes under the same address. MTR in UI mode merges them together (i think?), but that doesn't seem right since they are seperate routes.

Here's an example on jsdelivr.com response. Look at lines starting with h 9 and h 10. h represents host type, and numbers route index. The IP is duplicated. (ignore duplication in tags)
https://gist.github.com/patrykcieszkowski/d49d0574fa76ad5497d7eb42323c5365

I added a list of raw outputs to the measurement response, so we can debug it. We might delete or keep it later.

Metadata should round the numbers "jMin": 0.15600000000000003,

done.

jsdelivr/globalping-probe#58

patrykcieszkowski · 2022-06-15T10:45:18Z

Here's another example of this duplication on google.com domain. Except this time, one of them comes with a resolved hostname.

https://gist.github.com/patrykcieszkowski/9afa8a01456096ccfd27dc924e992424

jimaek · 2022-06-15T17:25:28Z

Another issue with private IPs when running mtr to google.com

{
  "id": "4rrYZPmmhTJo4sVf",
  "type": "mtr",
  "status": "finished",
  "createdAt": 1655313809031,
  "updatedAt": 1655313810260,
  "results": [
    {
      "probe": {
        "continent": "SA",
        "region": "southern america",
        "country": "EC",
        "state": null,
        "city": "guayaquil",
        "asn": 14522,
        "longitude": -79.8862,
        "latitude": -2.1962,
        "network": "satnet"
      },
      "result": {
        "hops": [],
        "rawOutput": "Private IP ranges are not allowed",
        "data": [
          "x 0 33000",
          "h 0 186.68.104.129",
          "p 0 575 33000",
          "x 1 33001",
          "h 1 10.145.72.1",
          "p 1 5864 33001",
          "x 2 33002",
          "h 2 192.168.0.162",
          "p 2 6178 33002"
        ]
      }
    }
  ]
}

{
  "id": "ovhiEtEyoWfz9NLo",
  "type": "mtr",
  "status": "finished",
  "createdAt": 1655313912181,
  "updatedAt": 1655313912733,
  "results": [
    {
      "probe": {
        "continent": "NA",
        "region": "northern america",
        "country": "US",
        "state": "VA",
        "city": "ashburn",
        "asn": 14618,
        "longitude": -77.4875,
        "latitude": 39.0437,
        "network": "amazon.com"
      },
      "result": {
        "hops": [],
        "rawOutput": "Private IP ranges are not allowed",
        "data": [
          "x 0 33000",
          "h 0 216.182.238.245",
          "p 0 1160 33000",
          "x 1 33001",
          "x 2 33002",
          "x 3 33003",
          "x 4 33004",
          "h 4 241.0.12.7",
          "p 4 373 33004"
        ]
      }
    }
  ]
}

jimaek · 2022-06-20T16:01:43Z

Current state:

Duplication bug when tests end with (waiting for reply)
Weird behavior when testing IPs like 1.1.1.1 and 8.8.8.8

jimaek · 2022-06-23T16:55:49Z

Some tests are randomly broken:

{
  "id": "qJYx6a573nNUGn1O",
  "type": "mtr",
  "status": "finished",
  "createdAt": 1656003058895,
  "updatedAt": 1656003059366,
  "results": [
    {
      "probe": {
        "continent": "EU",
        "region": "northern europe",
        "country": "FI",
        "state": null,
        "city": "helsinki",
        "asn": 24940,
        "longitude": 24.93,
        "latitude": 60.17,
        "network": "hetzner online gmbh"
      },
      "result": {
        "hops": [],
        "rawOutput": "Host                            Loss% Drop Rcv Avg  StDev  Javg \n1. AS???   _gateway (172.31.1.1)    0.0%    0   0 2.1    0.0   2.1\n2. AS24940 65.108.112.232 (65.108.112.232)    0.0%    0   0 0.3    0.0   0.3\n3. AS???   (waiting for reply) \n4. AS24940 88.198.244.241 (88.198.244.241)    0.0%    0   0 5.8    0.0   5.8\n5. AS24940 88.198.249.89 (88.198.249.89)    0.0%    0   0 0.6    0.0   0.6\n6. AS24940 213.239.224.37 (213.239.224.37)    0.0%    0   0 0.6    0.0   0.6\n7. AS???   (waiting for reply) \n",
        "data": []
      }
    }
  ]
}

Host                            Loss% Drop Rcv Avg  StDev  Javg 
1. AS???   _gateway (172.31.1.1)    0.0%    0   0 2.1    0.0   2.1
2. AS24940 65.108.112.232 (65.108.112.232)    0.0%    0   0 0.3    0.0   0.3
3. AS???   (waiting for reply) 
4. AS24940 88.198.244.241 (88.198.244.241)    0.0%    0   0 5.8    0.0   5.8
5. AS24940 88.198.249.89 (88.198.249.89)    0.0%    0   0 0.6    0.0   0.6
6. AS24940 213.239.224.37 (213.239.224.37)    0.0%    0   0 0.6    0.0   0.6
7. AS???   (waiting for reply)

Also need to test cases when the endpoint doesnt resolve correctly.

patrykcieszkowski · 2022-07-26T12:12:18Z

I believe this issue is now resolved. Have you noticed anything else wrong with MTR cmd?

jimaek · 2022-07-26T12:18:35Z

It looks good to me, but given the many edge-cases we encountered how about @MartinKolarik @ayuhito you also test our mtr?
To make sure it works properly. Maybe spend 10 minutes on it just in case

patrykcieszkowski · 2022-07-26T12:30:52Z

btw I don't think we've released any new probe version, and we changed a lot in the input, response schema etc. Maybe we should do it before testing anything.

jimaek · 2022-07-26T17:57:00Z

Everything was merged and deployed. You can now all test again MTR and maybe even the rest of the commands too to make sure the latest changes didn't break anything.

jimaek assigned patrykcieszkowski May 16, 2022

patrykcieszkowski mentioned this issue Jun 2, 2022

feat: mtr command #138

Merged

4 tasks

jimaek closed this as completed in #138 Jun 14, 2022

jimaek reopened this Jun 14, 2022

patrykcieszkowski mentioned this issue Jun 21, 2022

fix: mtr fail on ip target jsdelivr/globalping-probe#63

Merged

jimaek closed this as completed Jul 29, 2022

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add mtr #127

Add mtr #127

jimaek commented May 16, 2022 •

edited

Loading

patrykcieszkowski commented May 16, 2022

jimaek commented May 16, 2022

patrykcieszkowski commented May 16, 2022

jimaek commented May 16, 2022

patrykcieszkowski commented May 24, 2022 •

edited

Loading

jimaek commented May 24, 2022

patrykcieszkowski commented May 24, 2022

jimaek commented May 24, 2022

jimaek commented Jun 14, 2022

patrykcieszkowski commented Jun 15, 2022 •

edited

Loading

patrykcieszkowski commented Jun 15, 2022

jimaek commented Jun 15, 2022

jimaek commented Jun 20, 2022

jimaek commented Jun 23, 2022 •

edited

Loading

patrykcieszkowski commented Jul 26, 2022

jimaek commented Jul 26, 2022

patrykcieszkowski commented Jul 26, 2022

jimaek commented Jul 26, 2022

Add mtr #127

Add mtr #127

Comments

jimaek commented May 16, 2022 • edited Loading

patrykcieszkowski commented May 16, 2022

jimaek commented May 16, 2022

patrykcieszkowski commented May 16, 2022

jimaek commented May 16, 2022

patrykcieszkowski commented May 24, 2022 • edited Loading

jimaek commented May 24, 2022

patrykcieszkowski commented May 24, 2022

jimaek commented May 24, 2022

jimaek commented Jun 14, 2022

patrykcieszkowski commented Jun 15, 2022 • edited Loading

patrykcieszkowski commented Jun 15, 2022

jimaek commented Jun 15, 2022

jimaek commented Jun 20, 2022

jimaek commented Jun 23, 2022 • edited Loading

patrykcieszkowski commented Jul 26, 2022

jimaek commented Jul 26, 2022

patrykcieszkowski commented Jul 26, 2022

jimaek commented Jul 26, 2022

jimaek commented May 16, 2022 •

edited

Loading

patrykcieszkowski commented May 24, 2022 •

edited

Loading

patrykcieszkowski commented Jun 15, 2022 •

edited

Loading

jimaek commented Jun 23, 2022 •

edited

Loading