AI Chat

Master conversational server management with Hive's AI Chat interface.

Starting a Conversation

1Select Server - Click on a server in the Hive dashboard

2Open AI Chat - Click the AI Chat tab

3Ask Questions - Type naturally in plain English

Example Conversations

Investigating High CPU

You:"The website is slow and CPU is at 95%"

Hive: I'll investigate the high CPU usage.

✓ Running: ps aux --sort=-%cpu | head -20

Found the culprit:

PID USER CPU% COMMAND
12847 www 89% php-fpm: pool www

✓ Running: strace -p 12847 -c (sampling for 5 seconds)

The process is making excessive database calls. The /api/search endpoint is running an unoptimized query scanning 2 million rows.

Recommendations:

Add database index on products.name column
Implement pagination on search results
Consider caching frequent searches

Disk Space Emergency

You:"Server is out of disk space"

Hive: I'll find what's consuming the disk space.

✓ Checking disk usage... /dev/sda1 is 98% full

✓ Running: du -sh /* | sort -rh | head -10

Largest directories:

/var/log     → 45 GB (67% of used space)
/var/lib     → 12 GB
/home        → 8 GB

✓ Running: du -sh /var/log/* | sort -rh | head -10

Found: /var/log/nginx/access.log → 38 GB (single file!)

This nginx access log hasn't been rotated and grew to 38GB.

I can help you:

Safely truncate this log file
Set up log rotation to prevent this

You:"Yes, truncate it"

⚠️ This will permanently delete the log contents.

Command: truncate -s 0 /var/log/nginx/access.log

ApproveReject

Service Won't Start

You:"MySQL won't start"

Hive: I'll diagnose the MySQL startup issue.

✓ Checking service status...

mysql.service - MySQL Server
Status: failed (Result: exit-code)

✓ Checking recent logs: journalctl -u mysql -n 50

Error found:

[ERROR] InnoDB: Unable to lock ./ibdata1 error: 11

Another MySQL process is holding the lock file.

✓ Checking for running MySQL processes...

Found: Zombie MySQL process (PID 8834) still holding the lock

To fix this, I need to:

Kill the zombie process
Remove stale socket file
Start MySQL service

Kubernetes Pod Troubleshooting

You:"The payment service pods keep crashing"

Hive: I'll investigate the payment service pod crashes.

✓ Running: kubectl get pods -l app=payment-service

NAME READY STATUS RESTARTS
payment-service-7d4b8c9f6-x2k4m 0/1 CrashLoopBackOff 5

✓ Running: kubectl logs payment-service-7d4b8c9f6-x2k4m --previous

Error found:

FATAL: Unable to connect to database at db.internal:5432
Connection refused

✓ Running: kubectl get svc db -o wide

Found the issue: The service selector doesn't match the pod labels.
Service has selector "app: database" but pod has label "app: postgres"

Fix: Update the service selector to match the pod label.

You:"Yes, fix it"

⚠️ This will modify Kubernetes resources. Approval required.

Command: kubectl patch svc db -p '{"spec":{"selector":{"app":"postgres"}}}'

ApproveReject

Docker Container Investigation

You:"The redis container keeps dying"

Hive: I'll investigate the Redis container issue.

✓ Running: docker ps -a | grep redis

CONTAINER ID IMAGE STATUS NAMES
a3b4c5d6e7f8 redis:7 Exited (137) 2 minutes ago redis-cache

Exit code 137 indicates OOM kill. Checking logs...

✓ Running: docker logs redis-cache --tail 100

Out of memory trying to allocate 16384 bytes

✓ Running: docker inspect redis-cache --format='{{.HostConfig.Memory}}'

Root cause: Redis has 128MB memory limit but dataset exceeded this.

Recommendations:

Increase container memory limit to 512MB
Configure Redis maxmemory policy
Restart the container

You:"Yes, set memory to 512MB and restart"

⚠️ Approval required for container operations.

Commands: docker stop redis-cache && docker rm redis-cache && docker run -d --name redis-cache -m 512m redis:7

ApproveReject

Nginx 502 Bad Gateway

You:"Website is showing 502 bad gateway"

Hive: I'll diagnose the 502 Bad Gateway error.

✓ Checking nginx status... nginx is running (PID 1234)

✓ Running: tail -50 /var/log/nginx/error.log

[error] connect() failed (111: Connection refused) while connecting to upstream, upstream: "http://127.0.0.1:3000"

Nginx is trying to proxy to port 3000 but nothing is listening.

✓ Running: systemctl status node-app

Found: node-app service is stopped (failed)

✓ Running: journalctl -u node-app -n 50

Error: Cannot find module '/var/www/app/server.js'

Root cause: The Node.js app file /var/www/app/server.js is missing.
Last modified: 2 hours ago (possible bad deployment?)

Recommendations:

Check your deployment - server.js may not have been deployed
Roll back to previous version if available

Full-Stack Investigation

You:"Users are complaining the checkout is broken"

Hive: I'll investigate the checkout issue across the full stack.

✓ Step 1: Checking web server (nginx) → nginx is running, no errors

✓ Step 2: Checking application (Node.js)

Found: "PaymentService: Connection timeout to payment-api:8080"

✓ Step 3: Checking payment service → Connection refused!

✓ Step 4: Checking payment container → exited 23 minutes ago

✓ Step 5: Checking container logs

Error: Redis connection failed - max retries exceeded

✓ Step 6: Checking Redis → Over memory limit (2.1GB / 2GB)

ROOT CAUSE FOUND:

Redis hit memory limit → Payment service couldn't write →
Payment service crashed → Checkout API returns 500

Fix sequence:

Increase Redis memory limit to 4GB
Restart Redis container
Restart payment-api container

You:"Yes, fix it all"

✓ Redis memory increased to 4GB

✓ Redis restarted

✓ Payment service is healthy

✓ POST /api/checkout returned 200 OK

Checkout is now working!

Security Investigation

You:"Check if anyone unauthorized accessed this server"

Hive: I'll perform a security audit for unauthorized access.

✓ Checking authentication logs...

⚠️ Found 847 failed SSH login attempts in the last 24 hours

Top attacking IPs:

192.168.1.105 → 312 attempts (internal IP - concerning!)
45.33.32.156 → 245 attempts (external)
167.99.123.45 → 180 attempts (external)

✓ Checking successful logins...

Successful logins in last 7 days:

admin from 10.0.0.5 (3 times) ← Your office IP
deploy from 10.0.0.10 (12 times) ← CI/CD server

✓ No suspicious successful logins detected

✓ Checking for new user accounts...

✓ No new users created in last 30 days

Summary:

✓ No unauthorized access detected
⚠️ Brute force attempts ongoing (847 in 24h)
⚠️ Internal IP 192.168.1.105 needs investigation

Recommendations:

Install fail2ban to block brute force attempts
Investigate the internal IP
Consider disabling password auth (use keys only)

Capabilities

Hive helps with a wide range of DevOps, SRE, and system administration tasks. Use it to investigate issues, inspect systems, and assist with guided operational work.

Server Health & Monitoring

Task	What Hive Does
Health check	Checks CPU, memory, disk, services, network
Performance analysis	Identifies bottlenecks and resource hogs
Service status	Checks if critical services are running
Uptime report	Shows system uptime and recent reboots

Troubleshooting

Task	What Hive Does
Error investigation	Finds errors in logs, traces root causes
Service failures	Diagnoses why services won't start
Performance issues	Finds slow queries, memory leaks, CPU hogs
Connectivity problems	Tests network, DNS, ports, firewalls

Log Analysis

Task	What Hive Does
Find errors	Searches logs for errors and exceptions
Track patterns	Identifies recurring issues
Filter by time	Shows logs from specific time periods
Correlate events	Connects related events across log files

System Administration

Task	What Hive Does
Service management	Start, stop, restart services (with approval)
Process control	Find and manage running processes
File operations	Find, view, and manage files
Configuration	Check and update config files

Kubernetes & Containers

Task	What Hive Does
Pod management	Check pod status, logs, describe pods, restart pods
Deployments	View deployments, rollout status, scale replicas
Services & Ingress	Check service endpoints, ingress configurations
Docker containers	List containers, view logs, inspect, exec into containers
Container debugging	Investigate why containers crash, check exit codes

Web Servers & Proxies

Task	What Hive Does
Nginx/Apache	Check config syntax, view access/error logs, reload
SSL/TLS	Check certificate expiry, debug SSL issues
Load balancers	Verify upstream health, check configurations
Reverse proxy	Debug routing issues, check headers

Databases

Task	What Hive Does
MySQL/PostgreSQL	Check status, show slow queries, connection issues
Redis/MongoDB	Check memory usage, connection count, replication
Query analysis	Find expensive queries, check locks, deadlocks
Backup verification	Check backup status, verify dump files

Networking & Cloud

Task	What Hive Does
DNS debugging	Check resolution, dig, nslookup
Connectivity	Test ports, trace routes, check firewalls
Cloud CLIs	Run aws, gcloud, az commands for cloud resources
VPN/Tunnels	Check tunnel status, debug connectivity

Security

Task	What Hive Does
Login audit	Shows who logged in and when
Failed attempts	Identifies brute force attacks
Open ports	Lists network services and ports
File permissions	Checks for insecure permissions
Vulnerability scanning	Check for outdated packages, exposed services

Multi-Step Investigation

Hive automatically continues investigating until it finds the root cause:

Initial Check

Checks application logs → Finds "Connection refused" error

Follow the Trail

Connection to what? → Checks error details → It's the database

Check Database

Is MySQL running? → No! Service is stopped

Why Did It Stop?

Checks MySQL logs → Out of memory kill (OOM)

✓

Root Cause Found

Memory leak in background job processor consumed all RAM

Context Awareness

Hive remembers your conversation and understands context:

You:"Check the web server"

✓ Nginx is running, listening on ports 80 and 443

You:"What about the config?"

Hive understands you mean "nginx config"
Server blocks: 3, SSL: Enabled, Gzip: Enabled, Config test: ✓ OK

You:"Compare it to the database server"

Hive understands you want to compare nginx between servers
db-01 doesn't have nginx installed. Did you mean MySQL?

Approval Workflow

When Hive needs to make changes, you're always in control:

✓

Auto-Run

ls, ps, cat, grep, df, free, top, netstat

Needs Approval

restart, kill, install, config changes

Blocked

rm -rf /, shutdown, format

Learn more about Safety & Security.