Akamaistermetric trend

The Quiet Reality of Building - 3/11/2026 One thing that continues to surprise me is how quiet the response is to much of the work I'm publishing. Today I reflected again on the contrast between: • hundreds of articles written • a functioning software system • months of engineering effort versus the small number of reactions or comments online. But this may simply be the nature of building early. The internet rewards novelty, humor, and fast-moving memes. Infrastructure rarely attracts attention while it is being built. The real value only becomes obvious **later**. So the work continues.

Akamaistermetric trend

Private Relay as a Debugging Tool - 3/11/2026 Running a private relay locally has already proven useful for development. Even though the relay currently runs only on my machine, it provides several advantages: • deterministic testing • controlled publishing behavior • whitelisted writers • predictable event storage Unlike public relays — which sometimes fail, timeout, or behave inconsistently — a local relay behaves exactly as expected. For debugging complex publishing flows, this is extremely valuable. Eventually this relay could also be exposed via a reverse proxy as something like: wss://relay.continuum.xyz But for now, running it locally inside Docker is already useful.

Akamaistermetric trend

Backfilling a Local Relay - 3/11/2026 One experiment I'm considering is **backfilling my local relay**. I currently have roughly: • 700+ articles • 500+ notes If all of those events are republished to the local relay, the relay becomes a **complete local mirror** of my writing history. This would create a few advantages: • faster queries • deterministic storage • easier testing of relay behavior • simpler debugging The relay effectively becomes a **local indexing engine for Nostr events**. The main thing to watch is limits such as `maxLimit = 500` in strfry configuration. But since this limit only affects queries, not stored events, it should not prevent the relay from holding the entire archive. Again, something to test later.

Akamaistermetric trend

Relay Lists as Persistent Artifacts - 3/11/2026 Another architectural question surfaced today: **where should my relay list live?** Nostr has a built-in solution: kind:10002 — relay list metadata. Because I already publish relay lists, Continuum could simply retrieve the **latest 10002 event** and use that as the canonical relay configuration. This would mean: • relay configuration becomes portable • relay lists become part of identity history • new installations automatically inherit correct relay configuration Instead of storing relays only in local configuration files, they could become **signed identity artifacts**. That is a much more elegant model.

Akamaistermetric trend

Local Relay Backfill Experiments - 3/11/2026 Today I started thinking more seriously about using my local relay (running strfry in Docker) as a **performance optimization layer**. Right now Continuum can rebuild the archive from: 1. The deterministic archive repo 2. Multiple public relays 3. Local cached data But a **locally running relay** opens another interesting possibility. If I backfill the relay with all of my existing events (700+ articles and 500+ notes), the relay effectively becomes a **local event index**. That means: • startup reads could be faster • queries could be simpler • archive rebuilds might become optional The open question is performance. Is reading from a local relay faster than rebuilding from the archive? The only way to know is to test it. This isn't urgent, but it's an intriguing direction.