Watermark error when trying to access the canister via agent-js

Developers JavaScript
1

Constant error when using agent-js, canister inaccessible with 75% of calls failing with

Error while making call: Timestamp failed to pass the watermark after retrying the configured 3 times. We cannot guarantee the integrity of the response since it could be a replay attack.

https://dashboard.internetcomputer.org/canister/iwv6l-6iaaa-aaaal-ajjjq-cai

How can I solve this?..
Why is this keep happening?

2

Can you check your system time and make sure that it is not lagging or ahead?

1 Like
3

(adding for better discoverability)

There is a forum post with this error message where it was happening for different agent-js versions, additionally there was a suggestion to optionally disable watermark check (with corresponding PR).

I would encourage people who report this (or similar issues)

  • to provide their agent-js version
  • some minimal steps to reproduce the issue
  • maybe also to comment on their experience with retry strategy (if that helps or not)
  • add ‘+1’/thumbs-up or whatever emoji to the feature request PR to indicate you’re affected by the issue too
2 Likes
4

No, system time is set perfectly correct(windows, time sync enabled, tz correct)

Issue is kind of wide-spread, I’ve also been chatting with other dev yesterday and he seen a lot of this in his front-end as well
It leads to incredible sluggishness as part of requests never finish at all(absolutely ruins UX)

5

The gateway (not the IC itself) was under heavy load yesterday, I had issue accessing websites but, now everything is back to normal. Therefore I was wondering, are you still facing the issue today?

1 Like
6

“@dfinity/agent”: “^2.3.0”
default retry strategy I guess(3 attempts)

I’d rather see a proper solution that does not disable watermark checking(whatever this is) as I understand it can significantly lower the security

What is the root cause of this issue anyway?

7

I’ll try updating to the 2.4.1 and see if it still persist
But this is an ancient issue, I’ve been devving dapp for past half year with various agent versions and it’s been always here

8

I use 10 (#yolo :sweat_smile:) but according to the comment I left in the code as documentation, the error it resolves isn’t the same as the one you originally mentioned. :man_shrugging:

Not sure it helps in your case, maybe worth a try anyway?

1 Like
9

Just also put it to 10. Nice one…
At least requests hopefully won’t just fail now but just take longer

Any idea on what is the root cause and how this can be adressed? As far as I understand it’s an issue with nodes…
but why this keeps happening(sporadically)?

10

If the issue only happened yesterday, I think it was related to what I mentioned above.

If it starts happening more often, I have to admit I’m not entirely sure — hence the #yolo hashtag in my previous post :sweat_smile:

I remember initially using the option on one side due to heavy network load (few months ago), but also because it helped overcome a bug in agent-js that (I think) was fixed in v2.3.0. In the end, I just stuck with 10 retries because why not? But again, it might just also help avoid a different issue and not help in yours at all.

1 Like