Share your provider experience - help build a public Monad validator reliability picture

General
,
1

Hey validators :waving_hand:

For the past 6 months we have been tracking every provider-related incident affecting Monad validators in our tracker Monad Network Dashboard - Validators & Block Space | Pulse by Unity Nodes

We just published the first semi-annual overview based on that data:
:open_book: Full article: Monad Provider Reliability Report - Semi-Annual Overview | Unity Nodes Blog

Short version: the report covers 7 provider events, 2 security events, and the full concentration picture across TeraSwitch, OVH, Hetzner, Cherry Servers, MEVSPACE. The Hetzner Solana ban in 2022 and the TeraSwitch routing failure in August 2026 give serious context for why this matters right now. Especially considering that the upcoming VDP revamp will incentivise decentralization along ISP/ASN and geographic lines - meaning provider information becomes much more actionable.

Why this post:
We won’t catch every incident on our own. And provider reliability is exactly the kind of information every validator needs before deciding to run infrastructure on one host or another.

If you’re running a Monad validator (or have run one), please drop your experience in a reply. Attributed or fully anonymous - both work. All submissions will feed into the next quarterly report (Q4 2026, December) and become a public reference point for everyone else.

Example questions worth covering in your submission (free form is also fine):

  • Provider + datacenter (city, country) - which host, where exactly
  • How long have you been using this provider?
  • How many incidents in that period - and what kind (network / hardware / DC-wide / support delay)
  • Longest single downtime you experienced on this provider
  • Support responsiveness - how fast do they answer, do they solve it on the first try
  • Transparency - do they publish post-mortems after incidents, or not
  • Bandwidth and network stability
  • Price vs value
  • Would you recommend this provider to another validator?
  • Other providers you tried, and why you left

The more validators contribute, the more useful the rating becomes for the whole community. One submission = one solid data point that another operator will factor into their choice.

Thanks in advance :folded_hands:

1 Like
2

Happy to add ours. I run a validator on a dedicated OVH box in Sydney (Advance line, EPYC + NVMe). Bare metal has been the right call for me: the io_uring and disk path Monad leans on is sensitive to noisy neighbours, and a dedicated NVMe for the triedb kept block lag far more stable than anything shared would. The APAC location adds a little propagation distance to a mostly US/EU set, but it has been fine for keeping up.

One thing that might help what you’re building: a chunk of the reliability picture is already visible as data. I run MonadPulse (https://monadpulse.xyz), which tracks the live validator set with an ASN/provider and geo breakdown, so you can see where stake actually concentrates by provider rather than by self-report. Could be a useful cross-check against the experiences people post here. Glad to compare notes on specific providers too.

1 Like
3

Hey guys, here is my experience with MEVSPACE. this is based on one server in Poland so obviously just one data point

I used it for around 7 months in total. first 5 months were fine, then i started getting random network outages. it wasnt only ssh, during the best documented incident both ping and ssh were dead from outside while kvm and local logs showed the os and Monad were still running. the physical link was up but the server couldnt reach its gateway or the internet

I can document roughly 5-6 significant incidents, plus a few short network flaps. reboot usually restored access temporarily. the longest measured outage was 5h 44m, partly because i deliberately left the server in the failed state so support could investigate it live

eventually we captured NETDEV WATCHDOG transmit timeouts on the motherboard Realtek RTL8125 using the default r8169 driver. MEVSPACE also saw r8169 errors on the console. they suggested switching to the dedicated r8125 driver, but that helped for less than a day and the same problem came back

after that they offered an additional Intel NIC for €1 net per month, with testing before billing. the amount is tiny of course, but after spending a week troubleshooting the server myself it still felt a bit funny. also the first card they installed was actually Broadcom using tg3, not Intel. they apologized for the mix-up and then installed an Intel card using igb and moved the network connection to it

after that the server stayed stable for around 8 weeks. during the maintenance an unrelated BIOS setting was also reset and SMT was enabled again, so i had to ask them to restore my original no-SMT configuration

then the server became unreachable again. i waited around 30 minutes for support and finally rebooted it through the panel myself. access returned after the reboot, and only then support replied that the server was pinging. well yes, it was pinging because i had already rebooted it

I didnt do another full investigation that time so i cant say the final outage had the same root cause. honestly i just didnt want to spend another round diagnosing provider hardware myself and moved elsewhere

when the server was up i had no particular bandwidth complaints. the price was good, but for a production validator the downtime and time spent troubleshooting were not worth the savings for me. support was polite but meaningful action often required several follow-ups and a lot of evidence from my side. personally i wouldnt recommend it for production unless you are ready to troubleshoot these things yourself and can tolerate the downtime