Sudden unexplained drop in scrape_samples_scraped within every 24 hours

Hello,
I have a 4 services running, namely, Prometheus, Promscale, Postgres/Timescaledb and Grafana to monitor my django application.
I noticed that one of my metrics to count total API RPM get sudden dips and it deviates from what my elastic apm shows. So I decided to check the samples scraped at the time of dip and it matched. Whenever there is a sudden drop in samples scraped, there is drop in my metrics too.

Here is the image of samples scraped graph:

Docker-compose of my services:

version: “3.3”

services:

postgres-db:
image: Package dev_promscale_extension · GitHub
ports:
- 5432:5432/tcp
networks:
- rexnet
environment:
- POSTGRES_PASSWORD=***
- POSTGRES_USER=***
- POSTGRES_DB=***
deploy:
placement:
constraints: [node.labels.trex.type ==mon]

promscale:
image: timescale/promscale:latest
ports:
- 9201:9201/tcp
- 9202:9202/tcp
depends_on:
- postgres-db
- prometheus
networks:
- rexnet
environment:
- PROMSCALE_DB_CONNECT_RETRIES=10
- PROMSCALE_DB_HOST=postgres-db
- PROMSCALE_DB_PORT=5432
- PROMSCALE_DB_NAME=***
- PROMSCALE_DB_SSL_MODE=allow
- PROMSCALE_DB_USER=***
- PROMSCALE_DB_PASSWORD=***

deploy:
  placement:
    constraints: [node.labels.trex.type ==mon]

prometheus:
image: prom/prometheus:main
user: ***
ports:
- 9090:9090
networks:
- rexnet
command: --web.enable-lifecycle --config.file=/etc/prometheus/prometheus.yml
volumes:
- ./prometheus:/etc/prometheus
- prometheus-data:/prometheus
- /var/run/docker.sock:/var/run/docker.sock:ro
deploy:
placement:
constraints: [node.labels.trex.type ==mon]

grafana:
image: grafana/grafana:7.5.7
ports:
- 3000:3000
networks:
- rexnet
restart: unless-stopped
volumes:
- ./grafana:/etc/grafana/provisioning/datasources
- grafana-data:/var/lib/grafana
deploy:
placement:
constraints: [node.labels.trex.type == mon]

networks:
rexnet:
driver: overlay
attachable: true
external:
name: rexnet

volumes:
prometheus-data:
grafana-data:

Is there some particular issue is prometheus version or is it a resource allocation problem. I also noticed that when there is actual huge spike in my metrics on elastic apm, at that time there is this sudden dip in samples scraped.
Any help or suggestion is welcome.
Thank you.