January 2025 - ripe-atlas - mailman.ripe.net

v2 probe repair?
by Phillip Remaker 24 Nov '25

24 Nov '25

Hello! I have an 11+ year old v2 probe Seems to be in a reboot loop. Starts up, all LEDs go blank, and it repeats. Any diagnostics I can do? Any repairs? https://atlas.ripe.net/probes/4111/#tab-network

4 7

Re: Changes to RIPE Atlas API keys
by Ponikierski, Grzegorz 26 Feb '25

26 Feb '25

Hi! > * We'll deprecate and remove the ability to use the URL form in about a > year (around October 2024). > * We plan to send further reminders about this change over time, as well > as reaching out to heavy users of the to-be-removed format. It’s October 2024. Is this plan still valid? I admit that I have a bunch of tools which use URL form which I would have to update to use "Authorization" header. Regards, Grzegorz From: Robert Kisteleki <robert(a)ripe.net> Organisation: RIPE NCC Date: Tuesday, 19 September 2023 at 12:11 To: "ripe-atlas(a)ripe.net" <ripe-atlas(a)ripe.net> Subject: [atlas] Changes to RIPE Atlas API keys Dear RIPE Atlas users, We'd like to update you on some upcoming changes regarding API keys in RIPE Atlas. TL;DR nothing changes regarding how you can use your API keys in the short term - as long as you're actually using them. However, we'll change how unused or forgotten keys are handled as well as remove the less secure in-URL use of them. At the moment RIPE Atlas users can query their existing API keys via the UI and API, including the possibility to retrieve old keys. In order to improve the security of how we handle these, we'll introduce the following changes in October 2023: * The listing (retrieval) of keys will only reveal parts of the keys (enough to identify them) in the API as well as in the UI. * We'll add the ability to "regenerate" an API key, which will replace the secret UUID of the key while keeping exactly the same permissions. * Unused API keys will automatically be frozen after 1 year of not being used. Active keys (i.e. the ones that have been used at least once) will not be frozen. You still have the ability to save your keys until these changes are done and, as written above, you will be able to regenerate them later. We'll notify this list when the changes are about to be done. In addition, in order to further increase the security of our system, in the long run we'll make changes about how these API keys are communicated to the API: * At the moment the API accepts these either in HTTP headers ("Authorization" header) or in the URL (?key=xyz), although the Authorization header version has been documented as the preferred version for some time. * We'll deprecate and remove the ability to use the URL form in about a year (around October 2024). * We plan to send further reminders about this change over time, as well as reaching out to heavy users of the to-be-removed format. Regards, Robert Kisteleki RIPE Atlas team -- ripe-atlas mailing list ripe-atlas(a)ripe.net<mailto:ripe-atlas@ripe.net> https://urldefense.com/v3/__https://lists.ripe.net/mailman/listinfo/ripe-at…<https://urldefense.com/v3/__https:/lists.ripe.net/mailman/listinfo/ripe-atl…>

2 5

Why measurements using api are apt to fail while those using web interface can usually succeed?
by 孙乐童 02 Feb '25

02 Feb '25

Dear all, I have been attempting to use Atlas to launch traceroute measurements recently. Initially, I used ripe.atlas.cousteau to run the measurements in batches, keeping the total number under 10,000 to avoid exceeding the measurement limit. However, only 10% of the measurements returned successful results, while the other 90% showed the status "scheduling denied." Next, I tried launching only 12 measurements in batch, where each measurement contained one traceroute from a specific probe to a designated IPv4 destination. Out of these, 4 succeeded and the other 8 showed the status "specified". Following that, I launched the 12 measurements one by one with 2-second intervals, resulting in a success rate of 6 out of 12. However, when I increased the interval to 3 seconds, the success rate dropped to 4 out of 12. Interestingly, when I created the same measurements for the designated probes and destinations through the web interface, all of them were successful. Below is the Python code I used to launch the measurements: fromripe.atlas.cousteauimport ( Traceroute, AtlasSource, AtlasRequest, AtlasCreateRequest, AtlasResultsRequest, AtlasLatestRequest, ProbeRequest ) traceroute=Traceroute(af=4, target=dstip, description="traceroute subpref ip {} from {}{}".format(dstip, probe_type, probe_value), protocol='ICMP') if probe_type =='asn': probe_value =int(probe_value) source = AtlasSource(type=probe_type, value=probe_value, requested=1, tags={"include": ["system-ipv4-capable"]}) _key = ATLAS_API_KEY_2 for _ inrange(2): atlas_request = AtlasCreateRequest( start_time=datetime.utcnow() + timedelta(random.randint(0, 1)), key=_key, measurements=[traceroute], sources=[source], is_oneoff=True ) (is_success, response) = atlas_request.create() if is_success: print('create_task', dstip, is_success, response) msm_id = response['measurements'][0] return Could you kindly let me know if there might be an issue with my approach or provide any advice on how to improve this process? I would greatly appreciate any guidance. Thank you very much! Best regards, Sun Letong

4 7

Ripe Atlas Measurement Results and API Usage
by Huseyn Gambarov 30 Jan '25

30 Jan '25

Hello. I was implementing a mechanism to read the measurement results from the Ripe Atlas for many asynchronous measurements. In this process, I have faced multiple questions regarding the internal mechanisms of Ripe Atlas, and I would be thankful if you could help me clarify them. First, in the Ripe Atlas, I sometimes get “-1” in the result indicating 100% packet loss and sometimes very long RTTs. Can you tell me please what is the condition when Ripe gives up on a target and marks it as "-1”? Second, I have implemented asynchronous reading using web sockets and faced a rate limit on several concurrent web socket connections. The error message said: 'There are too many connections from this IP address. Try making multiple subscriptions on the same connection.’ Can you tell me please, if there is a limit on the number of subscriptions on the same connection? To solve the issue in the second question, I implemented a polling mechanism that periodically checks if the measurement is finished. In order to not overload the API, I attempted to retrieve multiple measurement results in a single call using the following API endpoint: GET /api/v2/measurements/{pk}/latest/ The documentation of that API endpoint states: "Get the most recent results from one or more measurements”. However, I need help understanding how to specify multiple measurement IDs in this endpoint, because all options I tried failed with HTTP status code 405 (Method Not Allowed). Could you please show me the correct way of using this API endpoint? Thank you! With regards, Huseyn Gambarov

3 2

Atlas UI: Redesigned Probe Detail Pages in Beta
by Stephen Suess 29 Jan '25

29 Jan '25

Hello Atlas friends, We’re excited to announce the early beta release of our redesigned probe detail pages in the Atlas UI. Over the past few months, we’ve been rethinking the structure and usability of these pages, and we’re sharing this early version to gather your feedback as we continue to develop and refine the design and features. Here’s how they are organised: - Overview: Probe information, map, traffic chart, links to measurements, and well-known target stats. - Network: IPv4/IPv6 overview and history. - Results: Grouped time-series charts for well-known targets and built-in measurements. - Manage: Tools for Sharing, Notifications, Transfer, SSH Key management, and Write-off. This release is very much a work in progress, and you can expect further changes and improvements as we iterate on the design and features. You can try it out by enabling the beta toggle in your preferences (https://atlas.ripe.net/prefs) We’d love for you to explore the new design and share your thoughts—your feedback will be instrumental in shaping the final version. For the full release notes you can go here: https://atlas.ripe.net/docs/releases/website/2025-01-28.html Thanks for being part of the community, and we look forward to hearing from you. Kind regards, Stephen Atlas UI

3 2

weird probe "outage"
by Sulev-Madis Silber 28 Jan '25

28 Jan '25

i just recently got message that my 65133 mox probe is offline according to status logs, it was offline for a ~week before this notice was sent out. notices are sent after 30min offline i didn't touch probe nor network nor do i know why that report was late or whether it even was really offline this is some recent issue maybe? or brand new issue? sos shows probe itself is up since i got it, 3+ months

1 0

Schroedinger's AAAA?
by Peter Eckel 28 Jan '25

28 Jan '25

I have a strane effect with probe 10293. Under "System Tags" I have both "Resolves AAAA Correctly" and "Resolves AAAA Incorrectly" listed. Apparently only one of the tags can be correct - or am I missing something? Best regards and seasonal wishes, Peter.

5 4

Oddity in "atlas-probe-traffic" measurement data/graph
by ripe＠nurfuerspam.de 28 Jan '25

28 Jan '25

Hi, Probes report the traffic volume they see on the device's/machine's interfaces in built-in measurement #9002 (on software probes, this needs to be enabled manually for privacy reasons). I use that to keep an eye on the behavior/health of the probe's host machine via the graphs generated by RIPEstat, which works well most of the time. Sometimes, however, there is an oddity in the data, namely extreme outliers. See attached graph, showing a peak of greater than 6.6 Gbps, when the "normal" traffic rarely exceeds 2Mbit/s. The underlying data shows an unusual high amount of bytes/packets sent/received for one instant that is pretty much impossible, here the data points around the anomaly (from [1]): [...] { "timestamp": 1736330734, "bytes_recv": 70881.58333333333, "bytes_sent": 63566.92222222222, "packets_recv": 737.8166666666667, "packets_sent": 700.6, "interfaces": [ "eth0" ] }, { "timestamp": 1736330921, "bytes_recv": 925866981.9732621, "bytes_sent": 820879052.6524065, "packets_recv": 8351935.347593583, "packets_sent": 8177239.866310161, "interfaces": [ "eth0" ] }, { "timestamp": 1736331101, "bytes_recv": 106008.85555555555, "bytes_sent": 88055.43333333333, "packets_recv": 1100.5944444444444, "packets_sent": 971.6666666666666, "interfaces": [ "eth0", "he-ipv6" ] }, [...] I think I've seen similar behavior on all of my software probes at one time or another, don't recall whether this also previously occured with my HW probe. So far, I have not been able to manually reproduce this, but it seems this is typically related to issues in communication with the controller, be it something on the controller side, like was more frequent/pronounced during the migration to the new controller infrastructure. Be it something on the probe side, e.g., temporary issue in network stack configuration. My current hypothesis is that when the probe cannot send off the measurement data, the accumulated data _for this particular measurement type_ is somehow getting corrupted, and the corrupted data is then sent once the communication issues are resolved, and data upload resumes. E.g., in this case, from the logs, it looks as if a reregistration with the controller was being done during minute 10:08, with the measurement data for #9002 taken just when the re-registration started. Not sure whether that is just coincidence, or whether there is a causal relation between the two. Anyone else seeing something similar, occasionally at least? Any idea what might be going on? (Have not gotten too far yet scouring through the probe/measurement code to see whether that yields any hints.) Thanks! R. [1] https://stat.ripe.net/data/atlas-probe-traffic/data.json?probe_id=1008486&m…

2 2

Small upcoming change in measurement metadata daily dumps
by Robert Kisteleki 28 Jan '25

28 Jan '25

Hello, We plan to make a small change in the measurement metadata daily dumps available at https://ftp.ripe.net/ripe/atlas/measurements/. Up until now these files contain the "current working set" which is defined as any measurement that runs right now or has been running in the past 14 days. (This also means that there's considerable overlap between today's and yesterday's dump.) We'll shorten this period to 7 days in early February. This has zero effect on the vast majority of our users. Those who periodically download and use these files with lower than 7 day frequency will need to update their scripts accordingly. Regards, Robert Kisteleki for RIPE Atlas team

1 0

Iso installation issue with RIPE Atlas anchor for LACNIC (uy-mvd-as28000)
by Elisa 07 Jan '25

07 Jan '25

Good morning, We received the ISO image to install the anchor but we encountered an issue. When we execute the "Install RIPE NCC Atlas Anchor" (see image) at some point the machine tries to resolve ks.atlas.ripe.net <http://ks.atlas.ripe.net/>, but it isn't connected to the network yet because to connect to the network we have to be inside the operating system, the installation process can't continue. We've been following instructions in: https://atlas.ripe.net/docs/howtos/installing-atlas-anchor.html and the machine has the following characteristics: 1 CPU type x86-64-v2 4 GB ram 50 GB disco Se carga la iso del disco local de Proxmox Can you help us move forward? Kind regards, -- Elisa Peirano *Analista de Datos I+D* R&D Data Analyst /www.lacnic.net <http://www.lacnic.net> Latin American and Caribbean Internet Addresses Registry/

2 2