Revamp and improve testing for Rococo <> Westend bridge

[svyatonik]

Since we are going to migrate to polkadot-sdk repo soon (hopefully soon after v1 deployment), let's start with preparing some scriptps + dashboards to test our code on live Rococo <> Westend bridge. We'll drop Rilato<>Millau soon, so the goal is to try to keep as most test cases from there as possible. Obviously, we need to migrate existing dashboards + alerts too

What we have now for Rococo <> Wococo and what shall be directly migrated to Rococo <> Westend:

• periodically send messages from R to W and in opposite direction;
• alerts and dashboards.

Which tests may be ported from Rialto <> Millau to live Rococo <> Westend:

• periodically send messages from R to W and in opposite direction (will be ported from Rococo <> Wococo);
• most of dashboards and alerts (will be ported from Rococo <> Wococo);
• periodically send large batch of messages from R to W and in opposite direction. We'll achieve that by utilizing the same script that we use now for sending single messages. So e.g. once per day it'll send a batch of messages

What we can not do on live Rococo <> Westend:

• send arbitrary messages (size and/or weight);
• assume that if something breaks and we see alerts, then it is the bridge issue. Rialto<>Millau is restarted daily, so we e.g. replenish all our accounts (message senders, relays, ...). Here we could break everything by simply not doing something on time;
• test bridge between GRANDPA chains.

What new can be tested with live Rococo <> Westend:

• test bridge between parachains. That's actually automatic

[svyatonik]

Ok, I totally forgot about the issue I had with polkadot-js-apps + a bunch of transactions. TLDR: it limits a number of transactions by the same signer to 1 per block. We can write our own js package, but I've proposed polkadot-js/tools#511 - hope we will be able to use it at our deployments.

In any case - unless we pause/kill our relayer, we won't be able to reach the outbound messages limit at the source BH, because it takes some time to submit tx, so in 6s we'll have a limited number of messages. But it should at least show that we are able to deliver messages in batches

[svyatonik]

need a new docker image: paritytech/scripts#620