HostingGuru
Managed Redis caches.
Redis is a top-level tab in your HostingGuru dashboard, alongside Services and Databases. Pick a plan, create a cache, copy the TLS connection URL, and use it from anywhere — your services, your laptop, an external app. Caches are ephemeral by design (no persistence) — perfect for caching, sessions, and queue backends.
From sign-up to a working rediss:// URL
1. Open the Redis tab
In the dashboard sidebar, click Redis. If you haven't subscribed yet, you'll see three plans side-by-side: Starter (free), Pro ($10/mo), and Scale ($25/mo). The plans gate how many caches you can run and how big each one is.
2. Pick a plan
Hit Activate Free Plan for Starter (no credit card) or Get Started on Pro/Scale to go through Stripe checkout. The plan attaches to your workspace; every member of the workspace can manage caches under it.
3. Create a cache
Click New Cache, enter a name (my-app-cache), confirm. The cache provisions in a few seconds. While it's provisioning the dashboard shows a yellow callout; once it's active, the connection details appear.
4. Copy & connect
The detail page shows the connection URL (masked, click the eye to reveal), host, port, and password. Connect with TLS enabled — use the rediss:// protocol prefix (not redis://). Most clients (ioredis, redis-py, go-redis) parse the URL directly.
Three plans, picked by cache count and memory
Starter — Free
1 Redis cache · 5 MB memory · 50 connections · TLS encrypted · ephemeral. The "kick the tires" plan — enough memory for a session store on a side project or to test out caching patterns. No credit card required.
Pro — $10/month · Popular
3 caches · 100 MB memory each · 500 connections · TLS encrypted · Pub/Sub support. The right plan for a single product with separate cache, session, and queue instances. Pub/Sub unlocks fan-out patterns (websocket broadcast, real-time notifications).
Scale — $25/month · Best Value
5 caches · 512 MB memory each · 2,000 connections · TLS encrypted · Pub/Sub support. The right plan for a team running several products, or one app with high cache pressure. 5× the memory of Pro for 2.5× the price.
How upgrades and downgrades work
Change plan at any time from Redis → Change Plan. Upgrades take effect immediately with prorated billing. Downgrades take effect immediately too — if you have more caches than the new plan allows, you'll be prompted to delete the excess first. Going from a paid plan back to free Starter requires going through Stripe checkout when you next upgrade.
A real Redis 7, not a sandbox
Redis 7 engine
Official Redis 7. Strings, lists, hashes, sets, sorted sets, streams, hyperloglog, geo, scripts. Pub/Sub on Pro and Scale. Standard Redis client protocol — every popular library works out of the box.
TLS by default
The connection URL uses rediss:// (not redis://). Encryption in transit is mandatory — there's no plaintext option. The auth password is included in the URL; the wire is encrypted end-to-end.
Ephemeral by design
Caches have no persistence. Memory is fast, simple, and predictable — ideal for caching, sessions, rate limiting, and queue backends where data is regenerable. Don't store data you can't reconstruct: when the cache restarts (rare, but possible during maintenance), it comes back empty.
Instant provisioning
From "Create cache" to active typically takes a few seconds. Status is shown as a badge — provisioning while it's coming up, running once ready. Connection details only appear when the cache is active.
Auth password masked & copyable
The detail page hides the password by default behind ••••••••••. Click the eye icon to reveal, the copy icon to copy. The same UX for the full connection URL. Easy to grab when configuring a service, hard to leak by accident on a shared screen.
Independent from services
Caches are workspace-level resources, not tied to a service. Use one cache from multiple services, share between API and worker, or use it from outside HostingGuru entirely. Deleting a service doesn't delete its cache.
What people put in Redis
HTTP response cache
Cache the result of expensive API calls or DB queries with a TTL. Easy 80% reduction in DB load for read-heavy endpoints. SET key value EX 60 patterns or any cache library.
Session store
express-session, django-redis-sessions, connect-redis — point them at your Redis URL and sessions persist across deploys, replicas, and machines. Logged-in users don't get bumped on every push.
Queue backend (Pro & Scale)
BullMQ, Sidekiq, Celery, RQ — they all use Redis as the broker. Pair with a worker service that consumes from the queue. Pub/Sub commands required for some queue libraries are available on Pro and Scale, not Starter.
Rate limiting
Token bucket per IP or per user via INCR + EXPIRE, or libraries like rate-limiter-flexible with the Redis store. Sub-millisecond per check, even at thousands of requests per second.
Pub/Sub fan-out (Pro & Scale)
PUBLISH from your API, SUBSCRIBE from a websocket worker. One server-side write, many connected clients receive. Pub/Sub is included on Pro and Scale.
Distributed lock
SET key value NX PX 10000 for a 10-second lock. Useful for "only one cron should run this minute" coordination. The single-instance simple SET NX pattern is enough — Redlock is overkill here.
Common questions
Is data persisted across restarts?
No. Caches are ephemeral — there's no AOF or RDB persistence. If your cache restarts (rare; only during maintenance), it comes back empty. Use Redis for data you can regenerate (cached query results, session tokens, queue jobs in flight). For durable data, use a managed Postgres database.
Why TLS-only?
The cache is reachable over the public internet. Plaintext Redis on the public internet is a security incident waiting to happen. TLS-only ensures the password and the cached data are encrypted in transit, with zero opt-out.
How do I connect — what's the URL format?
rediss://default:<password>@<host>:<port>. The rediss (with two s's) signals TLS to the client. Most clients parse this directly. If your client requires separate fields, the dashboard exposes host, port, and password individually.
Do connections count across all caches?
The connection limit is per cache, not per plan. Pro gets 500 connections per cache × 3 caches = 1,500 total. Scale gets 2,000 × 5 = 10,000 total. In practice, a well-tuned application uses connection pooling and stays under 50 connections per process.
Can I run multiple caches?
Yes — the plan limit is the count: 1 on Starter, 3 on Pro, 5 on Scale. Each cache is fully isolated with its own URL and password. Common pattern: separate cache and queue caches so eviction in the cache doesn't take down your queue.
What happens when I delete a cache?
Permanent — the dashboard requires you to type the cache name to confirm. All cached data is lost immediately. The plan limit goes back up by one. Stripe billing isn't affected (caches don't bill individually; the subscription is for the plan).
What if my subscription lapses?
If your Stripe subscription becomes inactive (failed card, canceled), the dashboard hides the connection details on the cache detail page and shows an amber "subscription not active" callout linking back to Billing. Re-subscribe to restore access.
Add Redis to your stack — start on the free Starter.
1 cache, 5 MB, TLS encrypted. No credit card. Upgrade when you outgrow it.
Key takeaways
HostingGuru Managed Redis is Redis 7.2 with TLS, RDB snapshots, and AOF on Pro plans — exposed as a REDIS_URL env var. Suitable as a queue backend, cache, session store, and Pub/Sub bus.
- Default eviction policy:
noeviction— safe for queues, no silently dropped jobs. - RDB snapshots every 60 seconds plus AOF every 1 second on Pro and above.
- Pub/Sub and Streams enabled by default, no per-channel limits.
- Internal connection from sibling services has sub-millisecond latency.
Should I use HostingGuru Redis as a cache or as a queue?
Both, but configure differently. For pure caching, switch eviction policy to allkeys-lru so the cache shrinks under pressure. For queues (BullMQ, Sidekiq, Celery), keep the default noeviction so jobs are never silently dropped.
How do I monitor HostingGuru Redis memory usage?
The dashboard surfaces memory used, peak memory, keys per database, and evicted keys (when applicable). Set up an alert on the Pro plan to be notified before you hit 80% of your plan's memory ceiling.