I ran into the same problem when first setting up a similar program stack. The problem I found was that some programs need to have the rclone mount setup before they first attempt to access the local folders. Just allowing the programs to start randomly would typically result in I/O errors.

My solution was to use the healthcheck feature of docker-compose to ensure that the mount was accessible before anything downstream booted. I placed a small "test" file in the mount and wait until it's available. I use bash, which means the rclone container needs to have it installed. Below is the dockerfile I use for rclone and my docker-compose file that controls the boot-up sequence.

One thing to note: only docker-compose uses the depends-on argument, so when your server boots-up the default is that docker starts the containers at random. You can get around this by turning off the auto-start using docker and then setting a cron-job that runs the docker-compose up commands at start-up.

Rclone Dockerfile FROM rclone/rclone:latest RUN apk update &&
apk add bash

docker-compose.yml version: '3.5'

services:

## rclone
    rclone-plex-media:
        build: ./rclone/
        container_name: rclone-plex-media
        environment:
          - RCLONE_CONFIG=/config/rclone.conf
        cap_add:
          - SYS_ADMIN
        privileged: true
        devices:
          - /dev/fuse
        restart: unless-stopped
        volumes:
          - ./rclone/plex-media:/mnt/docker/media/rclone/plex-media-internal:shared
          - ./rclone/plex-cache:/cache
          - ./rclone/plex-config:/config
        command: mount "plex-encrypted:"  "/mnt/docker/media/rclone/plex-media-internal" --uid 1000 --gid 1000 --umask 002 --default-permissions --allow-non-empty --allow-other --vfs-cache-mode full --cache-dir "/cache/" --vfs-cache-max-size 50G --dir-cache-time 1000h --vfs-cache-max-age 9999h --vfs-write-back 24h --log-level INFO --poll-interval=15s --stats 1m
        healthcheck:
          test: bash -c "[ -f /mnt/docker/media/rclone/plex-media-internal/test.txt ]"
          interval: 30s
          retries: 3
          start_period: 30s
          timeout: 5s
            
## Streaming - Plex
    plex:
        image: plexinc/pms-docker
        container_name: plex
        environment:
            - TZ=America/New_York
        devices:
            - /dev/dri:/dev/dri
        restart: unless-stopped
        network_mode: host
        volumes:
          - ./plex/config:/config
          - ./plex/transcode:/transcode
          - ./rclone/plex-media:/data
        depends_on:
          rclone-plex-media:
            condition: service_healthy
            
## Plex Scan Trigger - Autoscan
    autoscan:
        image: cloudb0x/autoscan
        container_name: autoscan
        environment:
          - PUID=1000
          - PGID=1000
        volumes:
          - ./autoscan/config:/config
          - ./rclone/plex-media:/mnt/unionfs:ro
        ports:
          - 3030:3030
        restart: unless-stopped
        depends_on:
          plex:
            condition: service_healthy        

## Media Management - Deluge
    deluge:
        image: binhex/arch-delugevpn
        container_name: deluge
        privileged: true
        sysctls:
          - net.ipv4.conf.all.src_valid_mark=1
        volumes:
          - ./deluge/downloads:/downloads
          - ./deluge/config:/config
          - /etc/localtime:/etc/localtime:ro
        environment:
          - VPN_ENABLED=yes
          - VPN_USER=REMOVED
          - VPN_PASS=REMOVED
          - VPN_PROV=REMOVED
          - VPN_CLIENT=wireguard
          - STRICT_PORT_FORWARD=yes
          - ENABLE_PRIVOXY=yes
          - LAN_NETWORK=192.168.2.0/24
          - NAME_SERVERS=REMOVED
          - DELUGE_DAEMON_LOG_LEVEL=info
          - DELUGE_WEB_LOG_LEVEL=info
          - VPN_INPUT_PORTS=1234
          - VPN_OUTPUT_PORTS=5678
          - DEBUG=false
          - UMASK=000
          - PUID=1000
          - PGID=1000
        ports:
          - 8112:8112
          - 8118:8118
          - 58846:58846
          - 58946:58946
        healthcheck:
          test: "curl -f http://localhost:8112 || exit 1"
          interval: 10s
          retries: 3
          start_period: 600s
          timeout: 5s
        restart: unless-stopped

## Media Management - Prowlarr
    prowlarr:
        image: lscr.io/linuxserver/prowlarr:develop
        container_name: prowlarr
        environment:
          - PUID=1000
          - PGID=1000
          - TZ=America/New_York
        volumes:
          - ./prowlarr/config:/config
        ports:
          - 9696:9696
        restart: unless-stopped
        healthcheck:
          test: "curl -f http://localhost:9696 || exit 1"
          interval: 10s
          retries: 6
          start_period: 5s
          timeout: 5s
        depends_on:
          rclone-plex-media:
            condition: service_healthy
          deluge:
            condition: service_healthy
    
## Media Management - Radarr
    radarr:
        image: lscr.io/linuxserver/radarr:latest
        container_name: radarr
        environment:
          - PUID=1000
          - PGID=1000
          - TZ=America/New_York
        volumes:
          - ./radarr/config:/config
          - ./rclone/plex-media/Movies:/movies
          - ./deluge/downloads:/downloads
        ports:
          - 7878:7878
        restart: unless-stopped
        healthcheck:
          test: "curl -f http://localhost:7878 || exit 1"
          interval: 10s
          retries: 6
          start_period: 5s
          timeout: 5s
        depends_on:
          rclone-plex-media:
            condition: service_healthy
          prowlarr:
            condition: service_healthy

## Media Management - Sonarr
    sonarr:
        image: lscr.io/linuxserver/sonarr:latest
        container_name: sonarr
        environment:
          - PUID=1000
          - PGID=1000
          - TZ=America/New_York
        volumes:
          - ./sonarr/config:/config
          - './rclone/plex-media/TV Shows:/tv'
          - ./deluge/downloads:/downloads
        ports:
          - 8989:8989
        restart: unless-stopped
        healthcheck:
          test: "curl -f http://localhost:8989 || exit 1"
          interval: 10s
          retries: 6
          start_period: 5s
          timeout: 5s
        depends_on:
          rclone-plex-media:
            condition: service_healthy
          prowlarr:
            condition: service_healthy

You're saying generic IO errors, but I've found rclone logs to be pretty specific when set to debug. At the very least what part of process do the errors start happening? Good place to start. Haven't used gdrive in a long time, though as I recall they used to have rate-limiting which could be a factor as well.

Another question is if the problem is moving after file completion, why stick to rclone if you're running into issues? Moving deluge completes to a local fs to check for errors in that process, and then either manual or a cron/timer/script with something like rsync for the final move would work for troubleshooting. At some point you're going to run into an error more detailed than "IO error."